JunTek PSG9080 Programmable Signal Generator (80 MHz - 300 MSa/s - 14-bit)

[pascal_sweden]

Anybody on this forum here who has experience with the JunTek PSG9080 Programmable Signal Generator?



Official website:
http://junteks.com/product/278017415

Pricing on AliExpress:
https://www.aliexpress.com/item/4001271854079.html?spm=2114.12010612.8148356.1.35d43705FtlC8A

[pascal_sweden]

Anybody with this signal generator?

[JOHN LEE]

I have already purchased one and will share it with you after receiving it

[pascal_sweden]

Excellent!

I look forward to your two upcoming videos on YouTube

1) Unboxing video

2) Product review video

[TheDefpom]

Juntek contacted me today asking if I was interested in taking a look at a unit, so if all goes well I will be getting one to do a review video on in the near future, so if you are not already subscribed to my YouTube channel I suggest you go there now and Sub to get to see the video when it is done (assuming all goes ahead).

https://www.youtube.com/TheDefpom

[pascal_sweden]

I contacted them before to ask them for a new unit so that I can do a review.

[electronx]

The price looks very good. Frankly, I wonder if I can send the mathematical functions through the communication port with the quick control interface matlab. Let's try it.

[JOHN LEE]

If you want to test PSG9080, you can contact their promotion person at "lk0629@foxmail.com". After the review is passed, he will send you a new product test. I hope my message will be helpful to you.

[SERJSOCHI]

Does this generator have a sweep sync output signal (see picture)?

[radiolistener]

Just got this unit. Front panel is beautiful and convenient in use.
Some buttons have glowing back-light, it looks very cool in the dark
Device case has a convenient stand and looks very good.

Output amplitude up to 25 Vpp (12.5 Vpk) below 1 MHz.


Minimum amplitude 1 mVpp (0.5 mVpk).


DC offset can be set from -10V to +14V with 0.01 V resolution.

Here is 11 MHz 6 Vpp with 50 Ohm oscilloscope input:


It has modulation mode from internal or external source.


Also it has a frequency meter and pulse counter mode for external source.According to my tests, the minimum sense voltage for external input is:
- 0.65 Vpk with AC coupling;
- 2.4 Vpk with DC coupling.

Let me know what you want to test

[TheDefpom]

I just finished recording the review and teardown for this (literally two minutes ago!)

The unit is pretty nice, it does have some output voltage linearity issues (it decreases as the frequency increases), but it doesn't really drop that much, you will be able to see that in the video when I publish it, as I show it and step through the frequencies to find out where its dips and peaks are.

I also really like the frequency fine tune in the menu which lets you adjust its output frequency to be more accurate, from factory it is at 50% adjustment, I adjusted mine to 58% to get it almost bang on frequency, unfortunately though this adjustment doesn't have an effect on the frequency counter input, that seems to be a seperate reference (or is not software tied to the fine tuning).

I will publish the video next week once I have had time to edit it etc.

[radiolistener]

I also really like the frequency fine tune in the menu which lets you adjust its output frequency to be more m from factory it is at 50% adjustment, I adjusted mine to 58% to get it almost bang on frequency, unfortunately though this adjustment doesn't have an effect on the frequency counter input, that seems to be a seperate reference (or is not software tied to the fine tuning).

yes, I also notice that frequency fine tune doesn't affect frequency meter.

Here is actual frequency output on my unit for 50 MHz setting and different frequency fine tuning value:
- 49.999713 MHz for setting 49
- 49.999765 MHz for setting 50 (default)
- 49.999817 MHz for setting 51
- 49.999869 MHz for setting 52
- 49.999921 MHz for setting 53
- 49.999973 MHz for setting 54
- 50.000026 MHz for setting 55
- 50.000078 MHz for setting 56

About 52 Hz per unit. 52*50 = ±2600 Hz tune range.
I think it's too high for usual TCXO, it will be better to have about 5 Hz per unit.

[TheDefpom]

The step size is dependant upon the output frequency, it is obviously a scaling factor/divisor rather than an offset adjustment, in my case it was only -12Hz off at 80MHz when set to 58.

[radiolistener]

Some teardown.

Front panel:
STM32F103RCT6
Crystal 8 МГц
LCD 3.4"

Frontend:
Xilinx Spartan-6 XC6SLX9
FPGA clock oscillator - 50 MHz
Relay HUiKE HK23F-DC5V-SHG
usb-serial CH340

Other IC's marking is scratched.
It looks like it uses R-2R DAC.

[radiolistener]

The step size is dependant upon the output frequency, it is obviously a scaling factor/divisor rather than an offset adjustment, in my case it was only -12Hz off at 80MHz when set to 58.

yes, I know, but it uses 50 MHz oscillator for FPGA clock.
So, when you set 50 MHz, the output will be exactly the same as FPGA oscillator frequency and you will see the real frequency error of hardware oscillator.

[TheDefpom]

Yep, I did a teardown video segment too, I noticed the same things as you with all the IC's lazered off apart from a couple.

I didn't identify the FPGA oscillator frequency as I couldn't read it at the time without getting it under a microscope.

[radiolistener]

here is fpga oscillator

[mawyatt]

That's a nice looking AWG and the interior looks well done!! Hopefully the waveform performance is also as good!!

Maybe those that have these can do some tests that show the waveform spectral purity and possibly a 2 tone test at a few frequencies?

Best,

[radiolistener]

Maybe those that have these can do some tests that show the waveform spectral purity

Here is spectrum measurement for sine 7 kHz 4 Vpp captured with PC sound card "line in" input at 192 kHz sample rate.
Main spectrum window RBW = 0.7 Hz.

The first picture is background noise of installation (sound card + cable connected to PSG9080, PSG9080 is physically disconnected from mains).


The second picture is 96 kHz span with second harmonic in the receiver bandwidth (2.7 kHz).


The third picture is zoom with first harmonic in the receiver bandwidth (2.7 kHz).

[TurboTom]

Thanks @radiolistener for the teardown and the tests. Considering the price, this "Junktek" thing is really disappointing. Cheap as cheap can...

No dedicated sample memory, R-2R discrete D/A converter, (probably) marginal reconstruction filtering and no shielding at all. The power supply , especially the transformer, doesn't look too trustworthy regarding creepage distances, and performance-wise, -40dBc for the second harmonic is really less than stellar!

This generator may be okay-ish for hobby applications, but if I had to spend 170+ EUR plus possibly import VAT/taxes, I'ld rather spend 60 EUR more (all taxes included) and get Rigol's DG811 and hack it. Despite the size and the toy-like appearance, this will get you a much more mature instrument which can be considered to be the entry into the professional segment. Have a look here for a set of teardown photos of my DG811. It becomes directly obvious that this generator plays in a different league.

If you don't want to spend that kind of money, the Feeltech generators are available for about half the money that the Junktek will set you back, and are pretty well characterized on this forum and, as the common opinion goes, are pretty useful as an entry level hobby AWG (with a mod applied or two...).

[TheDefpom]

I have a feeltech generator, it doesn't compare to this one when it comes to user interface and functionality, I would definately choose the PSG9080 over it, my impressions of it from having one here (my review video and teardown will be published on Wednesday) is certainly very different to yours, yes it has some limitations, but it also has some very nice aspects, the PCB's look well designed to me.

[mawyatt]

Maybe those that have these can do some tests that show the waveform spectral purity

Here is spectrum measurement for sine 7 kHz 4 Vpp captured with PC sound card "line in" input at 192 kHz sample rate.
Main spectrum window RBW = 0.7 Hz.

The first picture is background noise of installation (sound card + cable connected to PSG9080, PSG9080 is physically disconnected from mains).
(Attachment Link)

The second picture is 96 kHz span with second harmonic in the receiver bandwidth (2.7 kHz).
(Attachment Link)

The third picture is zoom with first harmonic in the receiver bandwidth (2.7 kHz).
(Attachment Link)

Was hoping for something better out of a 14 bit 300MSPS DAC at this low a frequency, which will likely get much worse as you go up in frequency.

Thanks for the testing.

Best,

[radiolistener]

It has external power supply connector 5V 3A, so it can be powered from external linear power supply. External power supply can improve performance.

when it comes to user interface and functionality

Yes, front panel of PSG9080 and user interface are excellent and very convenient in use. It is very comfortable to use it.

[mawyatt]

It has external power supply connector 5V 3A, so it can be powered from external linear power supply. External power supply can improve performance.

That's an interesting feature, one could power this from an external battery pack for portable use, however I doubt that an external PS will improved the output signal quality, easy enough to verify tho.

Best,

[JOHN LEE]

As far as I know, the maximum output frequency of DG811 is 10MHz, the sampling rate is 125Msa/S, and there is no programmable function, which can realize automatic output. I found that PSG9080 is grounded during use, which is safer to use. PSG9080 dual-channel maximum output is 80MHz. This price is quite cost-effective. If the requirements for AWG are not high, you can also consider JDS6600. I have a JDS6600 that has been working stably for 2 years. It works very well.

[TheDefpom]

My review video is online already for my Patreons, but I suppose I could give you guys a look at it before it becomes publicly available... here it is:

If you jump on to YouTube the video has links and chapters if you want to skip through it: https://youtu.be/b3yFxg2LwXI

[TurboTom]

As far as I know, the maximum output frequency of DG811 is 10MHz, the sampling rate is 125Msa/S, and there is no programmable function, which can realize automatic output. I found that PSG9080 is grounded during use, which is safer to use. PSG9080 dual-channel maximum output is 80MHz. This price is quite cost-effective. If the requirements for AWG are not high, you can also consider JDS6600. I have a JDS6600 that has been working stably for 2 years. It works very well.

That's why I suggested to hack the DG811 which basically turns it into a DG992 (less the top-end -- >70MHz -- level accuracy). For hobby use, this shouldn't cause much headache since it's completely reversible (in case of warranty claims) and is accurately documented in the corresponding thread. The DG800 is ground-referenced as well if that's a concern to you. I rather had it floating since this eliminates the risk of introducing ground-loops if working on a DUT or other test gear that's grounded itself.

Properly designed gear that's not ground-referenced isn't any more dangerous to use than properly designed grounded gear. And you can always ground-reference it by installing a grounding strap. The other direction isn't as safely possible to arrange.

But that's off-topic, sorry for that... 

[TheDefpom]

In the review I show how the earthing is setup in the unit (in the teardown section), you can choose whether or not to have it earth referenced.

[TurboTom]

Yes, in general that's a good approach. But having the ground wire lugs screwed down with self-tappers into plastic receptacles is more than a no-go. Or are there threaded brass inserts in the plastic standoffs of the lower casing member? It's plainly dangerous if there aren't.

Moreover, the BNC terminals are only attached to the PCB, there's no mechanical support at the front and rear panel. So every access to the BNCs will stress the main PCB.

This is all toy-like quality, IMO the instrument is too expensive for what you get.

[TheDefpom]

I think that electronically it is good, but mechanically there is some room for improvement, I would have preferred a metal chassis myself, for shielding and for the points you raised with regards to securing items to the case.

Even if they are using threaded inserts (I didn't remove any boards so I don't know) all plastics fail eventually and these will split (as a result of molecular shrinkage due to changes in the molecular weight, from the polymer chains breaking down into shorter ones giving increased density of the polymer chains, but this takes many years depending upon ambient temperatures, light exposure etc.).

[radiolistener]

According to my observations, there are 3 voltage bands with the following boundaries (it depends on voltage change direction - up or down):
- up 0.120/0.121 Vpp | down 0.100/0.101 Vpp
- up 0.361/0.362 Vpp | down 0.322/0.321 Vpp
- up 6.001/6.002 Vpp | down 5.602/5.601 Vpp

when you change voltage across these boundaries, some internal switch is changed.

Here is my measurement of output impedance for these voltage bands. Since 0.3-6 Vpp is the most usable band, I performed several measurements for this band. I used oscilloscope Siglent SDS1102X for amplitude measurement, so the result depends on oscilloscope measurement error.

Measurement was performed at 100 kHz frequency with two different loads - 50 Ohm and 1 MOhm. The first voltage value for 1 MOhm load and the second (after slash) for 50 Ohm load.

0.2960 / 0.1480 Vpp      =>  50.005 Ohm

0.992 / 0.504 Vpp      =>  48.418 Ohm
2.00 / 1.00 Vpp      =>  50.000 Ohm
2.96 / 1.49 Vpp      =>  49.333 Ohm
4.04 / 2.04 Vpp      =>  49.025 Ohm
4.92 / 2.44 Vpp      =>  50.825 Ohm
5.96 / 3.00 Vpp      =>  49.338 Ohm       

9.56 / 4.88 Vpp   =>  47.956 Ohm

As you can see, taking into account oscilloscope precision, the output impedance of PSG9080 is pretty close to 50 Ohm. Very good result

[electronx]

Do you hear capacitor noise when the device is off and the device is plugged in?
There is a thin capacitor noise when the device is turned off and the sound is gone when I turn it on. I can also hear the capacitor discharge sound when I unplug it while the device is turned off.

[radiolistener]

Do you hear capacitor noise when the device is off and the device is plugged in?

Yes, when I put my ear on the device case I can hear it.

[electronx]

frankly i don't think the power board is well designed. If we examine the mechanical design, it is really bad. The BNC sockets are not fixed, can't they take the devices of good brands as an example? As an electronics engineering student, I can do a much better mechanical design. I measure values in DC mode with a 5.5 digit multimeter. There are voltage differences such as 0.02 Volts. This is really suck. It has voltage fluctuations. It is good that it has an interface design and a large screen compared to its competitors.
I would definitely prefer this over Feelec. but if I had more money, I definitely wouldn't buy it. Even grounding in a plastic case is terrible

[radiolistener]

The BNC sockets are not fixed, can't they take the devices of good brands as an example?

from my opinion BNC connectors on front panel are good enough. They have good contact. Much better than cheap Chinese BNC that I seen on other devices.

[0culus]

Thanks @radiolistener for the teardown and the tests. Considering the price, this "Junktek" thing is really disappointing. Cheap as cheap can...

No dedicated sample memory, R-2R discrete D/A converter, (probably) marginal reconstruction filtering and no shielding at all. The power supply , especially the transformer, doesn't look too trustworthy regarding creepage distances, and performance-wise, -40dBc for the second harmonic is really less than stellar!

This generator may be okay-ish for hobby applications, but if I had to spend 170+ EUR plus possibly import VAT/taxes, I'ld rather spend 60 EUR more (all taxes included) and get Rigol's DG811 and hack it. Despite the size and the toy-like appearance, this will get you a much more mature instrument which can be considered to be the entry into the professional segment. Have a look here for a set of teardown photos of my DG811. It becomes directly obvious that this generator plays in a different league.

If you don't want to spend that kind of money, the Feeltech generators are available for about half the money that the Junktek will set you back, and are pretty well characterized on this forum and, as the common opinion goes, are pretty useful as an entry level hobby AWG (with a mod applied or two...).

Just looked at some US sellers of this fine instrument...over $200??? This thing is really a piece of junk. Might be worth $50. My recommendation to anyone looking at these is save your money and get an actual good quality signal generator from a reputable brand.

[JOHN LEE]

Regarding the function of SYNC, The manufacturer gave me a detailed explanation and took a picture today. I shared it with you. The SYNC function means the phase synchronization of the output signals of multiple machines. For example, when two machines are set to synchronize, when wiring The first machine is connected to SYNC-OUT, and the second machine is connected to SYNC-IN, then 4 signals can be output in phase.

[TheDefpom]

Just looked at some US sellers of this fine instrument...over $200??? This thing is really a piece of junk. Might be worth $50. My recommendation to anyone looking at these is save your money and get an actual good quality signal generator from a reputable brand.

What are you basing this revelation on? have you even used one, or seen one in person ?

I have used one, and I like it, it is a LOT better than some of the others I have tried, and for the money I actually think it is pretty good, it isn't just a function generator, or an ARB, but a SIGNAL GENERATOR.

Not everyone can afford a Keysight, or Rohde & Schwarz etc. that is why there are difference classes of instrument to suit different budgets, for the price class this is in it is excellent.

[electronx]

I used low-budget quality china products. I have one review here and it continues. https://www.eevblog.com/forum/testgear/tonghui-th1953-5-5-digit-benchtop-multimeter-hands-on-review/

I think tonghui is planning to launch a signal generator soon and I think it will launch a quality product for a low budget. and I will most likely buy it. It feels like it won't break even if I hit it with a hammer.
Obviously juntek  signal generator did not meet my expectations. Feelec -juntek and low-budget instruments like this are better than nothing. But when I contact the companies ,I see that they are only focused on making money. There is the idea that "if we make 100 dollars from everyone we'll get rich." Fpga s are expensive and I think that's the most money on this device. I think they make more than $ 100 for each product. Of course not everyone can buy rohde schwarz and keysight, but as I said there are much better china products.

[electronx]

voltage is very fluctuating. And when set to 15 volts, it appears 14.14 volts. The margin of error claimed by the manufacturer is 0.01 volts.
there is actually 6.08% margin of error. I measured without load, I can't imagine when I add load.
Of course, it is not good to calculate the percentage error for a value. However, even if I measure a lot and make average errors, the device does not meet the catalog values. It is clear that it cannot see 15 volts

[TurboTom]

Despite the mechanical deficiencies of this AWG, I wouldn't consider the offset range of +-15V to be the pinnacle to judge about its electrical capabilities. I reason that the output stage (since it apparently consists of a pair of opamps, presumably THS3091) is supplied by +-16V which would result in an output swing of just the +-14.2V measured at high impedance. There's a good reason that other manufacturers limit the high-Z output range of their AWGs to +-10V. They want to make sure that the output doesn't clip... So this may be an indication for bad firmware design but not necessarily for a general inaccuracy.

My approach to evaluate the instrument would be to have a look at the linearity of the DACs. Since they are of the discrete (individual resistors) R-2R type, tolerances in these resistors, especially those close to the MSB, will affect linearity, resulting in harmonic distortion. Moreover, since the R-2R-network is directly driven by the FPGA, all the inaccuracies and the noise superimposed to the FPGA supply rail that feeds the corresponding output bank, will add to the output signal. And then, the internal resistance of the FPGA output drivers adds to the DAC resistors as well. Alltogether, that's what results in a second harmonic of -40dBc as measured by @radiolistener. I'ld also like to see how clean the output is with no apparent signal present, i.e. using a ramp signal at an ultra-low frequency (microhertz) and then analyze the output spectrum. This will show how much noise is being coupled through the FPGA supply (which will probably depend on the momentary output level, i.e. if more of the FPGA outputs are driven high, there's more of the rail noise present than the other way round).

There's a good reason that A- and B-brand manufacturers spend the money on a "real" DAC which could be a considerable fraction of their BOM cost, next to the FPGA.

But to "play fair", I would limit all the testing to the output range that the other instruments can cover, i.e. +-10V High-Z or +-5V @ 50 Ohm load.

And then a look at phase jitter would be interesting. The "easy" way to check for that would be outputting a square wave of an "odd" frequency, triggering the scope on the signal and looking at it with a considerable delay. If the slopes are "smeared out", considerable jitter can be assumed to be present. For proper quantitative measurements, sophisticated frequency counters or a modulation domain analyzer (MDA, i.e. a HP 53310A) are better suited.

A side-by-side comparison against Rigol's entry level AWG which is in the same price range, using all this sophisticated gear, would be quite nice. But I won't be the one to spend the money on this generator, just to do the tests. I can anyway already imagine the results...  -- sorry. 

[0culus]

Just looked at some US sellers of this fine instrument...over $200??? This thing is really a piece of junk. Might be worth $50. My recommendation to anyone looking at these is save your money and get an actual good quality signal generator from a reputable brand.

What are you basing this revelation on? have you even used one, or seen one in person ?

I have used one, and I like it, it is a LOT better than some of the others I have tried, and for the money I actually think it is pretty good, it isn't just a function generator, or an ARB, but a SIGNAL GENERATOR.

Not everyone can afford a Keysight, or Rohde & Schwarz etc. that is why there are difference classes of instrument to suit different budgets, for the price class this is in it is excellent.

I once had something very similar, and it was a total piece of crap that ended up croaking for no obvious reason (probably the shitty SMPS, not worth fixing). It's not even heavy enough to be a door stop. Notice that I didn't exclude Rigol or Siglent equipment. There also used.

[JOHN LEE]

The maximum peak-to-peak value of the output signal of PSG9080 within 1MHz is 25Vpp, and the maximum forward peak value is 12.5V. We set the offset to 12.5V, and it can be seen that it is very accurate; in fact, it can be up to 14V ， there is no problem with the parameters.

[Johnny B Good]

 Nice review video as per usual, Scott.

 It was interesting to note the lack of anything that looked like a DAC. The resistors on each side of the FPGA look suspiciously like R2R ladder networks. It would be interesting to see what a full amplitude 100KHz sine wave looks like in detail at the half and quarter amplitude points. 

 I suspect your mains power arrangement (auto transformer and a UPS being mentioned) may be the reason why you're not seeing the full half mains live leakage on the BNC shields typical of such small class II smpsus, courtesy of the 1 or 2 nF Ycap that diverts the switching noise leakage between the HT and LV windings directly back to their origin in the direct rectified mains HT supply powering the switching chip in that 5v 3A psu board rather than let it take the scenic route and pollute the local environment. You might want to repeat that test whilst plugged directly into an ordinary mains outlet.

 I thought you'd made a rather valid point over taking advantage of the use of an unearthed class II psu to eliminate undesirable earth loops via a noise polluted mains grounding circuit, complete with random DC offsets and switching noise from other smpsu powered kit accompanying the primary 50/60Hz 'hum loop noise' so hated by the audiophool types.

 This was also a desirable feature of Feeltech's FY6600 and predecessor models, provided you never forget the "Ground first, unground last" connection sequence to protect susceptible DUTs from this ESD risk posed by that EMC mandated Ycap which curses all such class II smpsus. Happily, there is a simple way to have your cake and eat it in regard of this conflict of requirements and, funnily enough, the solution is a most superb example of a 'compromise' that gives more benefit than deficit. In this case it's as simple as connecting a 1 to 10 K 'static drain' resistor between the C14's PE tag circuit and the 0v rail of the main board.

 Assuming a 1nF Ycap and 240v 50Hz mains supply, a 10K 'static drain' resistor will knock this 120v (circa 90vac as seen with a typical DMM) leakage voltage down to a mere 377mVac (1K reduces it to 38mVac). If we assume a BNC cord shield resistance as high as one ohm, with a 'static drain' resistance as low as 1K being used to squash the half live mains leakage down to just 38mV or so, this will still provide some 60dB of attenuation to this unwanted ground loop noise pollution. You land up a winner all round - no ESD risk and no perceivable earth loop issue to contend with.

 Incidentally, it's worth pointing out that if you use a typical 5v 3A smpsu wallwart, you'll very likely have exactly the same half mains live voltage issue to contend with even if you've wired that half watt 'static drain' resistor to the PE pin on the C14 socket (unless of course, you've simply switched it off at the back and left it plugged into the mains).

 For anyone looking to expand their hobby interest in electronics and unsure about spending yet another three or four hundred quid on a 'mere signal generator' to go with a recently acquired DSO, these toy generators offer remarkably good value for the money. However, it seems these recent models' improvements over Feeltech's infamous FY6600 have been more by way of show than function as exemplified here with that waste of space front rubber 'bumper' that doesn't even look the part (it looks more like a hideous stage set dressing parody rather than the real thing as seen on mainstream brands).

 Whilst the improved 'haptics' are a long overdue improvement to the user interface (a keypad to directly enter parameter values), the basic electrical performance specs are none to little better that what Feeltech had been offering in their now two year old FY6600 models (and, in many cases, actually worse! ), so it's good to see these new contenders vying to displace Feeltech's products as the "Go To" cheap signal generator option being held up to close scrutiny like this.

 As for myself, after tinkering with my own FY6600 'tinker toy' for the past twenty months, I'm now rather looking forward to taking delivery of a nice shiny new SDG1032X this coming Monday morning 

John

[radiolistener]

just tested frequency meter in PSG9080, it works ok up to 245 MHz.

Unfortunately my si5351 cannot produce higher frequency and I don't have signal generator which can produce signal above 245 MHz. But I have UHF transceiver, so I tried to put UHF antenna into BNC connector of PSG9080 and tried to transmit 5W close to antenna. It works!

Frequency measurement of PSG9080 works up to 446 MHz!
And probably it may work with even higher frequency, but I don't have source to test it

[radiolistener]

It would be interesting to see what a full amplitude 100KHz sine wave looks like in detail at the half and quarter amplitude points. 

Here is 100 Hz 6 Vpp setting for sine and triangle waveform captured on 8-bit oscilloscope at 1 GS/s.

[radiolistener]

100 Hz 4 Vpp sine and triangle wave captured with sound card at 192 kHz

Also added zoom for triangle wave at zero crossing position.
And sine spectrum.

[maxwell3e10]

Having no external clock reference and 8192 waveform points puts it in the same class as FY6600

[Mechatrommer]

This generator may be okay-ish for hobby applications, but if I had to spend 170+ EUR plus possibly import VAT/taxes, I'ld rather spend 60 EUR more (all taxes included) and get Rigol's DG811 and hack it. Despite the size and the toy-like appearance, this will get you a much more mature instrument which can be considered to be the entry into the professional segment. Have a look here for a set of teardown photos of my DG811. It becomes directly obvious that this generator plays in a different league.

yes indeed different league... your DG811 is 20MHz @ $300. this Juntek is 80MHz @ $200... i guess to make apple to apple comparison with Juntek (or UTG962), Rigol DG4062 60MHz AWG ($800+) is better suited, imho...

Just looked at some US sellers of this fine instrument...over $200??? This thing is really a piece of junk. Might be worth $50. My recommendation to anyone looking at these is save your money and get an actual good quality signal generator from a reputable brand.

What are you basing this revelation on? have you even used one, or seen one in person ?

I have used one, and I like it, it is a LOT better than some of the others I have tried, and for the money I actually think it is pretty good, it isn't just a function generator, or an ARB, but a SIGNAL GENERATOR.

Not everyone can afford a Keysight, or Rohde & Schwarz etc. that is why there are difference classes of instrument to suit different budgets, for the price class this is in it is excellent.

I once had something very similar, and it was a total piece of crap that ended up croaking for no obvious reason (probably the shitty SMPS, not worth fixing). It's not even heavy enough to be a door stop. Notice that I didn't exclude Rigol or Siglent equipment. There also used.

can you recommend a good brand name that can go 60-80MHz at ~$200?

btw, here's flatness of my UTG962 (up to 60MHz) taken from some random project... (yellow synch signal, cyan sweep input (UTG962 output), deep blue is DUT sweep output purple NA). happy buy so far, i'm sorry for your experience, it happened sometimes...

[TurboTom]

yes indeed different league... your DG811 is 20MHz @ $300. this Juntek is 80MHz @ $200... i guess to make apple to apple comparison with Juntek (or UTG962), Rigol DG4062 60MHz AWG ($800+) is better suited, imho...
...

Sure. If you will use it as-is. I know of few to none private users or small businesses who won't consider hacking. Sorry, but IMO your point is ridiculous.

[Mechatrommer]

yes indeed different league... your DG811 is 20MHz @ $300. this Juntek is 80MHz @ $200... i guess to make apple to apple comparison with Juntek (or UTG962), Rigol DG4062 60MHz AWG ($800+) is better suited, imho...

Sure. If you will use it as-is. I know of few to none private users or small businesses who won't consider hacking. Sorry, but IMO your point is ridiculous.

how far can you hack that 125MSa/s device? i dont follow the hack thread sorry. i assume to the max 62.5MHz? do you call 2 points per cycle distortion free pro/business grade? not mentioning the other fun stuffs such as freq resolution when you go near that limit... enlighten me please.

[TurboTom]

The DG811 can be hacked to full DG992 performance (except for amplitude accuracy that drops by about 2dB between 70 and 100MHz -- unless we figure out how it's supposed to be calibrated). The hardware of the models is identical as a common service manual in chinese language, that's surfaced somewhere, proved (same P/N of main PCB). The information that the DG800 series has only 125MSa/s of scan rate is wrong and may have been published in a very early design phase by Rigol. All these units run at 250MSa/s.

Price-wise, I only compared the PSG9080 with the DG811 since in my part of the woods, provided I'ld had to pay full import duties and VAT for the JunTek unit, they would be on par (give or take ten eurobucks). Construction-wise, there is no comparison. Reliability and safety-wise, I trust Rigol's instrument 100% while the PSG9080 could, well, use some improvement...

I agree with you that as a hobbyist's low-cost AWG, the UTG962 appears to be a decent choice.

[Mechatrommer]

The DG811 can be hacked to full DG992

well unfortunately.. its not easily available for sale anymore (ebay)...

[TurboTom]

The DG811 can be hacked to full DG992

well unfortunately.. its not easily available for sale anymore (ebay)...

Really? In Europe, several T/M gear distributors have stock available. Btw, once again, the 125MSa/s specified on the linked page is wrong. All these units sample at 250MSa/s.

Or did you mean to refer to the UTG962 which really appears to be difficult to find meanwhile?

[Mechatrommer]

The DG811 can be hacked to full DG992

well unfortunately.. its not easily available for sale anymore (ebay)...

Really? In Europe, several T/M gear distributors have stock available. Btw, once again, the 125MSa/s specified on the linked page is wrong. All these units sample at 250MSa/s.

Or did you mean to refer to the UTG962 which really appears to be difficult to find meanwhile?

UTG962/932 is newly released sale is not stable yet i guess it sometime come and go... but checking, its currently available in aliexpress (china market, where else?) and many sellers start to selling it there, one is $139 https://de.aliexpress.com/item/4000573960327.html?spm=a2g0o.productlist.0.0.5d833d05WH9iPe&algo_pvid=9a579168-5b1e-4282-88f2-d993b612bae3&algo_expid=9a579168-5b1e-4282-88f2-d993b612bae3-1&btsid=0bb0623016012863212357791e36d5&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

batronix... is euro currency, and its not very friendly to us... well i stand corrected... dg811 also sold in aliexpress $279... https://de.aliexpress.com/item/10000098463394.html?spm=a2g0o.productlist.0.0.c6a635c7RYMVym&algo_pvid=b308c759-b632-44d5-838e-bb0bbecadc8d&algo_expid=b308c759-b632-44d5-838e-bb0bbecadc8d-1&btsid=0bb0624016012866100348760ec1be&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

i was not aware it can be hacked to 100MHz so i stand corrected, but still $200+ is something for me cheers.

[SERJSOCHI]

Hi. I received this generator today. The speaker is very loud, sealed it with tape, now it sounds much better.
What is the best way to remove the gray bumper, as it interferes with placing the device on the shelf.

[radiolistener]

Hi. I received this generator today. The speaker is very loud, sealed it with tape, now it sounds much better.

Nice trick, I think I will do the same, because speaker is really too loud.

What is the best way to remove the gray bumper, as it interferes with placing the device on the shelf.

It is very useful, it allows to put generator on the top of oscilloscope and it don't falling off with help of that rubber bumper.

[jonathanlemoine]

Unless I'm missing something, this doesn't have a 50ohm output. The voltage only lines up with the scope on a 1M input impedance.

[radiolistener]

Unless I'm missing something, this doesn't have a 50ohm output. The voltage only lines up with the scope on a 1M input impedance.

No. I measured the output impedance, it has pretty precise 50 Ohm impedance for all voltage ranges up to 6 Vpp.

I didn't measured output impedance for amplitudes above 6 Vpp, because I'm worry about my RF dummy loads (because they are low power and may be burned out from high power ). But I think it's output impedance the same 50 Ohm up to 25 Vpp.

If you're talking about display value, yes it always displays open output amplitude for internal 50 Ohm impedance. So, you're needs to recalculate it for your load. Since PSG9080 hardware output is 50 Ohm, for 50 Ohm load you can just divide displayed value by 2 to get Vpp (peak-to-peak) amplitude on the load. Or divide displayed value by 4 to get Vpk (peak) amplitude on the load.

Since it displays amplitude for open output, you're needs to apply correction factor for displayed amplitude in order to get Vpp value on specific load:
- for 50 Ohm load Vpp = displayed Vpp * 0.5
- for 75 Ohm load Vpp = displayed Vpp * 0.6
- for 600 Ohm load Vpp = displayed Vpp * 0.923077
- for 1 MOhm load Vpp = displayed Vpp * 0.99995
- for 10 MOhm load Vpp = displayed Vpp * 0.999995

You can look my output impedance measurements here. It is performed with Siglent SDS1102X oscilloscope. With taking into account oscilloscope measurement error, we can see that PSG9080 has pretty precise output impedance 50 Ohm.

PS: just tested impedance with multimeter Brymen BM867:

500 Hz square wave, 1.000 Vpp setting on the generator.
10 MOhm input = 499.76 mV RMS
49.80975 Ohm input (10 MOhm in parallel with [49.77 Ohm + 0.04 Ohm wires]) = 249.20 mV RMS

Calculated output impedance = 50.08 Ohm

[grumss]

I have just purchased one myself (as a toss-around sig gen)

It certainly has "character" but over all its surprisingly handy for its cost.

A few have mentioned that the frequency fine tune doesnt work with the frequency counter... I find that it does but just doesnt align with the generator section! Seems somebody at Junktek has made a "boo-boo" or fudged one of the sections to make things work.. maybe will be fixed with firmware?

Frequency counter sensitivity is low BUT it can also handle 20Vpp according to the manual! - be careful with this one!
Id love to see a amplitude "dbm" button- but can sort of understand why it isnt there.

Overall (so far)- i would recommend- as long ones expectation isnt too high    (it is what it is)

[radiolistener]

A few have mentioned that the frequency fine tune doesnt work with the frequency counter... I find that it does but just doesnt align with the generator section!

what is your firmware/fpga versions?

My one:

Hardware 1.20
Firmware 1.20
FPGA 1.20

[radiolistener]

Frequency counter sensitivity is low BUT it can also handle 20Vpp according to the manual! - be careful with this one!

There is ambiguity for MOD input specification. The manual claims that it has 0...3 V range. But it is used for VCO mode and the same manual claims that VCO voltage range is 0V-5V. So, it's not clear what is voltage range for MOD input.

I tried to test it with 0-5 V and according to ADC values, it looks that MOD input range is 0-5V, but it can see difference even for 5.2 V and above, with non linear code.

[grumss]

A few have mentioned that the frequency fine tune doesnt work with the frequency counter... I find that it does but just doesnt align with the generator section!

what is your firmware/fpga versions?

My one:

Hardware 1.20
Firmware 1.20
FPGA 1.20

Hardware 1.20
Firmware 1.20

Hardware version (on PCB) 1.30

[grumss]

I have just noticed on mine that channel 1 output becomes VERY noisy when i turn on modulation. May have to send it back

[radiolistener]

I have just noticed on mine that channel 1 output becomes VERY noisy when i turn on modulation. May have to send it back

what settings did you used?

[grumss]

I dont remember the settings for this screen shot, but it was noisy like this with all settings

See attachment for channel2 (which is fine)

[radiolistener]

Which frequency did you used for the carrier (for failed screenshot)?

[grumss]

The modulation was 1k /100% and i tested with the carrier frequency from 1Mhz all the way to 80Mhz... Think these screen shots were at a carrier frequency of 10Mhz

[radiolistener]

just tested on my one, both channels works ok (screenshots from the first channel, 10 MHz carrier)

[grumss]

Yep- thats how mine should look as well, i suspect i have a bad solder joint on one of the components... Anyway i have only had it a week or so, im not going to open it up and fix it unless i have to..

I dont think it is typical of the generator, I have just got one that wasnt QCed well enough..

Going through all the "have you followed the instructions?" stuff with the ebay seller as we speak.
(can understand this- but I AM NO BEGINNER

VK3XDK (used to design microwave stuff ... still an easy search on google)

[masterx81]

Interesting device, but i really not like the r-2r DAC 

[radiolistener]

Anyway i have only had it a week or so, im not going to open it up and fix it unless i have to..

You can open and check the PCB. May be there is some flux or solder.
There are just 4 screws and no seal labels.

[gwideman]

In case  it's of interest to readers: 

There's a Koolertron CJDS98 which appears to be the same as this Juntek 9080.

If you download the zip file software and docs package pointed to by the Koolertron vendor on amazon.com, it contains a manual titled PSG9080, and showing a picture of the Juntek.

Of course "same" is relative, as there could well be hardware and firmware variations.

[gwideman]

It has external power supply connector 5V 3A, so it can be powered from external linear power supply. External power supply can improve performance.

That's an interesting feature, one could power this from an external battery pack for portable use, however I doubt that an external PS will improved the output signal quality, easy enough to verify tho.

Best,

A significant merit of running it from battery would be to completely avoid ground loops, or the necessity to provide some current path from the sig gen ground to mains earth to avoid half-mains-voltage appearing on the BNCs (albeit at high impedance), as others have discussed.

It's of at least some note that the power supply strategy has changed between the older FY9600/9700 design. Those older ones had a power supply board that produced +/-13.5V and +5V, which were a bit of a pain to replace (see SDG Electronics YouTube series, if interested). 

But this Juntech PSG 9080 has the main PSU board responsible only for providing, one assumes, 5V (thus facilitating the external 5V supply feature), and the other needed voltages are generated on the main DDS+amplifier board.  So that would make replacement of the internal PSU board pretty easy, for those who might want to do that.

Graham

[radiolistener]

yes, you can use USB power bank to power it.
But note it consume about 0.7 A for low voltage output and up to 2 A for 20 Vpp output

[jpshea]

Can the PSG9080 do (constant frequency) AM and frequency sweep at the same time?

[radiolistener]

Can the PSG9080 do (constant frequency) AM and frequency sweep at the same time?

No, modulation mode and sweep mode are different modes and available on separate screens.

I'm wondering, what you're trying to do with AM modulation in sweep mode? 

[XeRoX]

Hello!!

I just beginner in electronics.

I also bought JUNTEK PSG9080 in DEC 2020. I wanna pass out triangle signal to op amp but with external voltage offset by another op amp (both op amps are powering by external 9V BAT->3.3V LDO->OP AMPS). Then I connect GEN_GND (just GND without output !) to my offset amp then oscilloscope start display offset voltage with some noise component. My offset op amp produce 100 mV signal. If I try power generator from classical supply with 50 Hz transformer can it help me keep my 100 mV signal in clear?

Also I have another question about this generator - can I use it for getting amplitude-frequency response diagrams ? I saw sweep mode but can't synchronize gen output with "in X" on oscilloscope(( Does it really possible with such equipment?


Thanks all for answers !!

[jpshea]

@radiolistener for EMC immunity testing

[ozcar]

Also I have another question about this generator - can I use it for getting amplitude-frequency response diagrams ? I saw sweep mode but can't synchronize gen output with "in X" on oscilloscope(( Does it really possible with such equipment?

If the sweep does not provide a sync output, maybe instead you could FM modulate one channel by the other.

[XeRoX]

Hello!! My question about noise is not correct and actual more. I just used oscilloscope incorrectly. I connect generator out both to osc and op amp input. Then I try to get signals from generator and from op amp input synchronously I just connect offset amp output to GND by generator and oscilloscope cabels    Sorry for incorrect question.... Also I already have checked and working chassis of power supply 5V / 3A with linear stabilizer) I'm not sure that I need in such construstion in generator right now ......

[radiolistener]

I'm not sure that I need in such construstion in generator right now ......

if you want to put external voltage offset, you're needs to remove DC offset of generator.
Otherwise you can burn out generator with external voltage.

It means you're needs to connect output through some capacitor or transformer or use some buffer amplifier.

[DROBNJAK]

I also really like the frequency fine tune in the menu which lets you adjust its output frequency to be more m from factory it is at 50% adjustment, I adjusted mine to 58% to get it almost bang on frequency, unfortunately though this adjustment doesn't have an effect on the frequency counter input, that seems to be a seperate reference (or is not software tied to the fine tuning).

yes, I also notice that frequency fine tune doesn't affect frequency meter.

Here is actual frequency output on my unit for 50 MHz setting and different frequency fine tuning value:
- 49.999713 MHz for setting 49
- 49.999765 MHz for setting 50 (default)
- 49.999817 MHz for setting 51
- 49.999869 MHz for setting 52
- 49.999921 MHz for setting 53
- 49.999973 MHz for setting 54
- 50.000026 MHz for setting 55
- 50.000078 MHz for setting 56

About 52 Hz per unit. 52*50 = ±2600 Hz tune range.
I think it's too high for usual TCXO, it will be better to have about 5 Hz per unit.

I have an application where I need the accuracy of 1 Hz or better at around 60-70 MHz.

Can you please tell me if you can achieve such accuracy with PSG9080 sig-gen?

What kind of signal generator can guarantee such high accuracy?

[mawyatt]

Doubt the PSG9080 or most lower cost AWGs can achieve that level of accuracy. Likely require an external reference like a GPS reference for use with the AWGs that have an external reference (10MHz) input.

Think most of the mid-level AWGs from Siglent and Rigol will work with an external GPS reference, I know our SDG2042X works with 10MHz external references although we don't have a specific GPS reference. The higher end AWGs from KS and Tektronix should also work with an external GPS reference.

Best,

[DROBNJAK]

Doubt the PSG9080 or most lower cost AWGs can achieve that level of accuracy. Likely require an external reference like a GPS reference for use with the AWGs that have an external reference (10MHz) input.

Think most of the mid-level AWGs from Siglent and Rigol will work with an external GPS reference, I know our SDG2042X works with 10MHz external references although we don't have a specific GPS reference. The higher end AWGs from KS and Tektronix should also work with an external GPS reference.

Best,

Thanks for the tip. I didn't know that GPS reference is so good. That can save me some money.

On another note, in Rigol's DSG800 1.5 GHz spec sheet is says, pg.2 section: "Internal Reference Frequency", that it can deliver "< 5 ppb" frequency accuracy. That's 5 parts-per-billion which I find astonishing. Its just pitty it is out of my price range.

Is < 5 ppb frequency accuracy really common in that price range of AWGs?

https://www.rigol-uk.co.uk/jg/wp-content/uploads/2020/06/Rigol-DSG800-Datasheet.pdf

[rf-loop]

Doubt the PSG9080 or most lower cost AWGs can achieve that level of accuracy. Likely require an external reference like a GPS reference for use with the AWGs that have an external reference (10MHz) input.

Think most of the mid-level AWGs from Siglent and Rigol will work with an external GPS reference, I know our SDG2042X works with 10MHz external references although we don't have a specific GPS reference. The higher end AWGs from KS and Tektronix should also work with an external GPS reference.

Best,

Thanks for the tip. I didn't know that GPS reference is so good. That can save me some money.

On another note, in Rigol's DSG800 1.5 GHz spec sheet is says, pg.2 section: "Internal Reference Frequency", that it can deliver "< 5 ppb" frequency accuracy. That's 5 parts-per-billion which I find astonishing. Its just pitty it is out of my price range.

Is < 5 ppb frequency accuracy really common in that price range of AWGs?

https://www.rigol-uk.co.uk/jg/wp-content/uploads/2020/06/Rigol-DSG800-Datasheet.pdf

This is only if you buy it with Optional OCXO reference.

Reference frequency 10 MHz
Temperature stability
In temperature range 0℃ to 50℃ , reference to 25℃ < 2 ppm
Aging rate < 1 ppm/year
With option OCXO-B08 < 5 ppb
With option OCXO-B08 < 30 ppb/year

+/- 5ppb is not very good.
+/- 50mHz @10MHz 

Also did not find what is initial accuracy after factory product line calibration. Hard to believe they can promise its initial limit is <5 ppb

standard version 2ppm alone mean max +/- 140Hz error with 70MHz and then also aging.

With Rb or GPS you can get lot of better than this OCXO Option version "5ppb" class what is "only"  5x10^-9 + aging etc.

[makoshatu]

Hi! Could anybody properly install the PC software for this PSG9080? I downloaded both the english and chinese version of the software from the Juntek website, but the zipped english version has CRC error. The CN version can be installed, however it could not register an .ocx controler. I added that .ocx file to the Windows and I registered it manually, so the software now can be run (an error message still show, but it can be okayed). Unfortunatelly, it does not show any text on the GUI, only the different values (frequency, amplitude, phase, etc.) can be set (if you know that what are you setting actually...).

[hflorin]

Hi there,

@SERJSOCHI

In your first post, you asked this:

Does this generator have a sweep sync output signal (see picture)?

    * sweep_sync.png (92.71 kB, 800x480 - viewed 533 times.)

In the second post I saw that you also bought one PSG9080:

Hi. I received this generator today. The speaker is very loud, sealed it with tape, now it sounds much better.
What is the best way to remove the gray bumper, as it interferes with placing the device on the shelf.

Please tell me, does this generator have a sweep sync output signal like in your picture?

Many thanks,
Florin

[radiolistener]

Please tell me, does this generator have a sweep sync output signal like in your picture?

Unfortunately no, it doesn't have sweep sync output.
Hardware allows to implement it, so I hope manufacturer will add such option in the future firmware.

[Neper]

Hi all,

I'm new here. Just got a new PSG 9080 yesterday. The brand of mine is JOY-IT instead of Juntek, but it looks identical.

Been playing around a few hours and so far things look OK. I have two short questions/remarks, if I may...

1. The Apple store has an App from Juntek for controlling a PSG9080 per Bluetooth from a iOS device.  Installed and tried it but it doesn't find the generator. JOY-IT say their PSG9080 doesn't have Bluetooth. Then again, there is no mention of Bluetooth in the Juntek docs either. Has anyone ever tried this App with the original Juntek 9080?

2. Is there a way of setting the output voltage in other units than Vpp? I'd much rather use Veff or dB.

3. JOY-IT have a firmware update for 'their' PSG9080 from March 10 on their website. Unfortunately no further info or version history. Just saying...

Ralf

[Northerner]

Hi Ralf,
1. You need a Bluetooth module to plug into the TTL port. Seems to be bundled by some vendors but not Joy-IT or Juntek.

2. AFAIK Vpp is the only way

3. The firmware seems to be v1.2 so no newer than what you probably already have.

[Neper]

1. You need a Bluetooth module to plug into the TTL port. Seems to be bundled by some vendors but not Joy-IT or Juntek.

The TTL port? Any idea if this would be avaible from somewhere else? Where does your photo come from? The funny thing is that the iOS software comes from Juntek.

2. AFAIK Vpp is the only way

How really stupid.

3. The firmware seems to be v1.2 so no newer than what you probably already have.

I've sent them an email and they say they've corrected one bug and one spelling error. I'm suitably impressed.

Ralf

[radiolistener]

2. Is there a way of setting the output voltage in other units than Vpp? I'd much rather use Veff or dB.

no, there is no way to change voltage units. Note, that PSG9080 shows Vpp for open output connector (with no load)!

When you connect some device with 50 Ω port (for example receiver or RF voltmeter with 50 Ω input), the actual Vpp voltage will be twice smaller than displayed on the PSG9080 display. Just divide displayed Vpp voltage by 2 and you will get actual Vpp voltage on the load. This is a bit annoying, but for some reason Chinese manufacturers don't want to use more common voltage units. For example, as I know FY6600 has the same issue.

For non 50 Ω load you will needs to recalculate actual voltage (the same as with any other signal generator).

3. JOY-IT have a firmware update for 'their' PSG9080 from March 10 on their website. Unfortunately no further info or version history. Just saying...

thanks for the link, it looks that they decided to rename brand name for some reason.

I don't recommend to use their software for PC. Because virustotal report it as dangerous.
And I don't see the reason for the PC software, because it's much better to control it from the front panel.

The only feature that needs PC software is to update firmware or upload arbitrary waveform. But if you decide to do it, it's better to use isolated PC with no internet. Do not share files from that PC with another computers for safety purposes.

I'm thinking about to write open source PC software to upload custom waveforms and firmware. But it needs protocol documentation and research. Unfortunately there is lack of information about protocol.



After some period of PSG9080 usage, I can report some found issues.


Firmware bugs (HW v1.20, FW v1.20, FPGA v1.20 - the latest FW at the moment):

1) Duty cycle don't works properly for the frequency 1'000'000 Hz and above. If you set 999'999 Hz or below that it works ok. But when you set frequency just 1 Hz above, the duty cycle setting is no more works as expected. This is obviously firmware bug, probably just some kind of mistyping in the source code or something like that. But be aware, with current FPGA firmware 1.20 you will be unable to setup duty cycle for the frequency 1 MHz and above. You can setup duty cycle for higher frequency on the display UI, but the actual signal output will be invalid. It stays close to 50% for any setting from 0 to 99.99%, but it is not exactly 50%, the actual duty cycle changes for about 2-5% instead of 0-100%.

2) FM modulation works incorrectly. When you connect FM receiver and test it, you will note a crackle sounds. There is some kind of overflow or incorrect signal processing mistake in the FPGA firmware 1.20. So you will be unable to use proper FM modulation with current firmware 1.20. For comparison AM and PM modulation don't have such issue.

Some pros and cons of PSG9080:

Pros:
- very nice front panel design and feel with convenient buttons and knob (it is based on STM32F103RCT6)
- large contrast LCD display
- FPGA module based on SPARTAN-6 XC6SLX9 which can be nice for custom hobby experiments, but there is a lack of documentation
- you can power PSG9080 from a uni-polar 5V source. For example I'm using 10 Ah USB power bank with 5V/2A output to get full galvanic isolation from mains (actual current consumption is about 0.7-1.2 Amps)

Cons:
- no output for a sweep sync
- sometimes it can produce overshoot voltage when you change settings. It needs to be used very careful with sensitive electronics
- very small SMD VCXO, so I'm afraid you will be unable to replace onboard oscillator
- about -40..-45 dBc even harmonics, probably due to R2R DAC.
- onboard DC/DC is not enough to keep proper voltage for 25 Vpp signals for any frequency and waveform type. If you use more than 10-12 Vpp, you're needs to check the actual output waveform on the load. I think it's possible to replace power supply with a good bipolar linear one.

Second harmonic is about -40..-45 dBc (sine waveform).
Actual bandwidth is about 60 MHz, it produce pretty nice sine up to 80 MHz, but above 60 MHz it has max 3.6 Vpp (max 1.3 Vpp on 50 Ω load) and some amplitude drop if you sweep from 1 MHz to 80 MHz.
Up to 30-40 MHz it has pretty nice flat response.

In overall, it works good enough and very comfortable for a lab experiments where you don't need a low noise or low harmonics distortions.


You can see some waveforms from PSG9080. I used a high speed 14-bit ADC AD6645 working at 96 MHz to capture it. For example, the first picture is a 10 kHz saw tooth waveform, you can see about 30% of the full wave height on the screen (I zoomed it to see more details).

[radiolistener]

just made command line tool for PSG9080 to upload and download arbitrary waveforms.

Will be available on github soon. 

[radiolistener]

Here is beta version 

For Windows XP and Vista it may need to install .NET Framework Runtime 4.0, for other Windows it should run just out of the box.

Usage:

Read arb wave from PSG9080 to a text file:

Code: [Select]

PSG9080_ARB.exe COM17 -read 1 wave01.txt
Write arb wave from text file to PSG9080:

Code: [Select]

PSG9080_ARB.exe COM17 -write 1 wave01.txt
COM17 is a serial port with PSG9080.
1 is arb wave number
wave01.txt is a file name

Note, it works with unsigned int 14-bit values in decimal format. So, the value should be within 0...16383 range.
The file should consists of 8192 values in text decimal format.

PSG9080 Generator software use 16-bit format for text file. It converts values before upload. And you never know which value will be written. This is why I implemented direct 14-bit format to avoid unexpected conversion loss. PSG9080_ARB uploads exactly the same data as it represented in the file. With no conversion.

In order to support PSG9080 software files with arbitrary wave in 16-bit format, I added two additional option. So the tool can work with text files in 16-bit format:

Read arb wave from PSG9080 to a text file with 16-bit format:

Code: [Select]

PSG9080_ARB.exe COM17 -read16 1 wave01.txt
Write arb wave from text file with 16-bit format to PSG9080:

Code: [Select]

PSG9080_ARB.exe COM17 -write16 1 wave01.txt

The tool has some data checks, so if something is wrong with the data, it will stop operation and show the error.

Also I implemented upgrade commands to upgrade firmware from file without installing a bunch of software, but I didn't tested it, so upgrade command is disabled in this build.

Source code will be published later on github.

[Neper]

My 9080 is too big to put it where I need it. All because of those stupid ears. I've had a closer look and they appear to be glued to the rest. Has anyone succeded in removing them without breaking everything?

Ralf

[radiolistener]

The source code for the arbitrary wave upload and download command line tool for PSG9080

https://github.com/qrp73/PSG9080_ARB

[torch]

Some questions for those that have bought one:

1. I read that one channel cannot modulate while the other is sweeping. Can one channel generate a basic waveform while the other channel is sweeping?

2. Is there a marker function?

3. The on-line manual shows trigger input and output bnc connectors on the rear, but from what I read here they don't appear to actually be present and the teardown video shows unpopulated points on the circuit board in that location. Are there different versions available? Did anyone get one with trigger I/O ports on the rear? (The JoyIt version also shows these rear panel connectors on the website images).

[radiolistener]

1. I read that one channel cannot modulate while the other is sweeping. Can one channel generate a basic waveform while the other channel is sweeping?

Yes, the second channel just keeps the last wave settings in Sweep mode.

2. Is there a marker function?

No.

3. The on-line manual shows trigger input and output bnc connectors on the rear, but from what I read here they don't appear to actually be present and the teardown video shows unpopulated points on the circuit board in that location. Are there different versions available? Did anyone get one with trigger I/O ports on the rear? (The JoyIt version also shows these rear panel connectors on the website images).

The rear BNC are used to connect with other PSG9080, one connector for previous PSG9080 and the second connector for the next PSG9080 in a sequence. It uses some internal and strange pulse structure, so you will be unable to use it as a trigger or a sync for the sweep.

[radiolistener]

some pictures from my tests taken with 14-bit ADC running at 96 MHz:

1) spectrum sine wave 3 MHz 4.1 Vpp (actual 2.05 Vpp at 50 Ω input)
2) spectrum square wave 3 MHz 4.1 Vpp (actual 2.05 Vpp at 50 Ω input)
3) oscillogram square wave 50 kHz
4) oscillogram ramp wave 10 kHz
5) oscillogram triangle wave 20 kHz

[torch]

Are those images of one of the channel outputs, or the strange output from the rear connector?

[radiolistener]

Are those images of one of the channel outputs, or the strange output from the rear connector?

pictures above are taken from the output channel. It was captured with a high speed/high dynamic range ADC to show signal purity details.

Here is rear pannel SYNC OUT connector output waves (yellow = CH1 output, magenta = rear panel SYNC OUT output).
First two pictures in normal mode square wave 20 MHz.
Pictures 3 is for a sweep mode from 10 to 30 MHz with 1 sec.

[radiolistener]

Does anyone tried to write custom opensource bare metal firmware for PSG9080?

I can try, but I'm afraid of losing the original firmware
Any idea on how to save it?

May be there is a sense to identify where FPGA firmware is stored, backup it and then try to desolder original STM32, replace it with a new one for experiments. And in case of needs just solder original STM32 and restore FPGA firmware. Or maybe just use custom STM32 devkit board instead of original front panel?

[Neper]

Just a short report from someone who's had his PSG9080 in the Joy-IT incarnation for a few weeks now...

I've been used to mostly surplus Eastern European lab gear of which ginormous quantities swamped the West German market after the fall of the Berlin wall. Now, that was 30 years ago and although all that stuff is still working fine (all of it is built like the proverbial battleship) I thought it might be time for something a little more modern, so I got hold of a PSG 9080 to replace a Soviet G4-102-A RF generator and an East German Praecitronic GF22 AF generator, both sinewave only.

The PSG9080 is giving me quite some brain gymnastics. I was used to my older AF generator that (a) is calibrated in Vrms and that (b) will output the voltage you set it to, no matter the load. You practically have to short-circuit it to make the output voltage drop below the set level. All my docs and instructions for the audio equipment that I work with ask for levels in Vrms. So, with the PSG9080 I find myself having to do a lot of mental acrobatics to convert forth and back between Vpp and Vrms and on top of that have to factor in the load resistance. Guess I'll keep my trusted old AF millivoltmeter connected in permanence to the PSG9080's output and set output levels watching the meter.

But the real prob I'm having with the PSG9080 is its noise, supposedly from the built-in SMPS. Not only on the output signal but also backwards via the mains into all the other equipment of my workshop. So, I replaced the original mains socket of the PSG9080 with one with a built-in mains filter. That got rid of some of the noise coming out at the mains side only to add more of it to the output signal. Running the PSG9080 from a linear (i.e. non-switching) 5 V power supply cured most of that. But I'm still somewhat shocked about how noisy that thing is. How are other people coping with this?

Oh, and I'm still looking for a way to get rid of those stupid ears without breaking the rest. They appear to be glued solidly to the housing. Not for cosmetic reasons, mind you, but to be able to fit in where I want it to go.

And I'll certainly be keeping my old friends from the East. :-)

Ralf

[radiolistener]

yeah, it's a little bit noisy due to switching mode power supply, but I didn't note noise interference for mains. May be because I plug it into power strip which has mains filter. You can power off it with mains switch on the rear panel.

Does anybody tried to replace switching mode PSU with linear bipolar PSU? I think it's very possible, but there is needs to research power and DC/DC circuit schematics on the analog frontend board to find where is analog ±voltage needs to be applied and where is digital voltage. When you power it from external 5V PSU, it still use DC/DC switching mode converters to produce ±voltage for analog amplifiers. I think with linear PSU it will be much better.

[fjalvingh]

I just received my PSG9080 and after seeing some odd behavior I decided to update its software... Bad plan; after the update (which itself said all went well) the device is dead: black screen.. I am trying to communicate with the AliExpress shop I got it from to get it back to life but that is... hard.....

It appears the device is updated through a serial connection from the USB port. That serial port is still there, but the thing listening to it seems to be dead.
Has anyone any idea whether I could update the firmware through some other means on that device?

[radiolistener]

I just received my PSG9080 and after seeing some odd behavior I decided to update its software... Bad plan; after the update (which itself said all went well) the device is dead: black screen.. I am trying to communicate with the AliExpress shop I got it from to get it back to life but that is... hard.....

It appears the device is updated through a serial connection from the USB port. That serial port is still there, but the thing listening to it seems to be dead.
Has anyone any idea whether I could update the firmware through some other means on that device?

Please clarify, the following questions:

1) what is device logo on your device? Juntek or JoyIT?

2) which flashing tool you're used (I know at least two tools available on internet), where did you downloaded it?

3) which firmware update file did you used? Where did you download it? What is MD5 checksum of the file?

4) do I understand correctly that flashing tool cannot see the device and there is no bootloader screen on the device after power on? I mean physical power on with mains switch on the back panel, because power on button on the front panel is software and don't works in bootloader mode.

Such information will help to solve your issue and also help other users to avoid such troubles.


I didn't tried to update firmware, because I don't have backup firmware image to restore device with JTAG. But I research firmware update tool with com port simulator and even tried to play with bootloader commands on device, but didn't tried to upload blocks So I know a little about bootloader and even wrote my firmware update code, but never tested it.

Since update is not success, your device should be still in bootloader mode. Even if you power off it from mains, it should enter into bootloader mode after power on.

The bad news here is that there is no full firmware update (MCU file and FPGA file). FPGA firmware update is available on Juntek site. MCU firmware update is available on JoyIT site. But probably they are not compatible with each other, so it needs to be careful and upload JoyIT firmware to device with JoyIT logo and Juntek firmware to device with Juntek logo. But I'm not sure, just didn't tried it.

You can identify firmware type by last two bytes in the file (in hex):
12 34 = FPGA update
56 78 = MCU update

[radiolistener]

As fjalvingh told me in PM, he used Juntek firmware update from this link:
http://68.168.132.244/PSG9080%20Firmware%20upgrade.zip

This is FPGA firmware update from the official Juntek download site.

I got the firmware and the update tool from an URL that was posted on Juntek's website and also on the device. It was
http://68.168.132.244/PSG9080%20Firmware%20upgrade.zip

The upgrade itself worked without any problem: the software did the update and said that it was a success. The device at that time also told me it was successful and that I needed to restart it by pressing a button. After pressing that button it restarted - and was dead. Dead meaning: black screen, and not even the software power button (the fron one with the blue light) reacts anymore: it stays on regardless of whether it is pressed or not.

My brand for the device is really "JUNTEK".

I hope I used the correct firmware, of course. I did see multiple versions on the nets but the one above was supposed to be the correct one. I also remember seeing that same URL (with the ip address instead of a host name) on the device itself - but I might be wrong there.

A complete power off/on only switches off the blue light in the power button and powering on lets the button burn again. Pressing it does nothing. The screen remains black. Connecting it to the PC makes the PC recognize a serial port, but running the upgrade again reports that the device does not react.

This is a good sign, it looks like the STM32 firmware is okay, but something is wrong with FPGA firmware. So the STM32 waits for FPGA response, but it don't works as expected. Probably this FPGA firmware update is not compatible with MCU firmware 1.20. I think it is stored on the frontend board in the flash memory IC. So probably it can be copied from a working PSG9080.

The problem is that we don't know which IC is a flash memory for XILINX Spartan 6. Any idea?

Another possible solution is to try to use FPGA JTAG connector to read/write the flash with FPGA firmware. I'm not familiar with XILINX family, I'm worked with Altera/Intel, but I think the XILINX chips should be similar and it should have standard flash memory interface which can be downloaded/uploaded through JTAG connector. And it looks that this 5x2 connector at the bottom of the photo is JTAG.

PS: PCB photo can be found here: https://www.eevblog.com/forum/testgear/juntek-psg9080-programmable-signal-generator-(80-mhz-300-msas-14-bit)/msg3234946/#msg3234946

[Hg7wgs]

Hi,

Two days ago arrived it, so yesterday night i try everything until 01.30 and today i wake up 6.30 after go work...
Now i find the solution!!!

Changed "decimal symbol" from "," to "." in windows(control panel, region settings, addictional settings, decimal symbol)

Hi! Could anybody properly install the PC software for this PSG9080? I downloaded both the english and chinese version of the software from the Juntek website, but the zipped english version has CRC error. The CN version can be installed, however it could not register an .ocx controler. I added that .ocx file to the Windows and I registered it manually, so the software now can be run (an error message still show, but it can be okayed). Unfortunatelly, it does not show any text on the GUI, only the different values (frequency, amplitude, phase, etc.) can be set (if you know that what are you setting actually...).

[radiolistener]

Two days ago arrived it, so yesterday night i try everything until 01.30 and today i wake up 6.30 after go work...
Now i find the solution!!!

Changed "decimal symbol" from "," to "." in windows(control panel, region settings, addictional settings, decimal symbol)

for what reason you're needs PC software?

[Hg7wgs]

Very easy drawing(and uploading) arbitrary waveform.

Two days ago arrived it, so yesterday night i try everything until 01.30 and today i wake up 6.30 after go work...
Now i find the solution!!!

Changed "decimal symbol" from "," to "." in windows(control panel, region settings, addictional settings, decimal symbol)

for what reason you're needs PC software?

[fjalvingh]

Hi radiolistener,

So far still no change. I am not sure that it is the FPGA that is the trouble. To me it sounds odd that you would first do a lot of FPGA work before you would show the splash screen. Also, the update program ITSELF fails telling "Failed or not responding" immediately. I assume the update is handled by the stm32, so to me it feels like that one is buggered.
I did get a new firmware file from our Chinese friends. This one contains an update ending in 56 78 so it should be an mcu update. But it cannot be placed due to the firmware upgrade issue.
There is an empty socket for a pin header on the display board; any idea whether that would be a jtag thing? Or is there any other way I could get that stm32 into boot loader mode or something similar?

[radiolistener]

I am not sure that it is the FPGA that is the trouble. To me it sounds odd that you would first do a lot of FPGA work before you would show the splash screen.

Your device is bricked. The most likely reason in such case is that MCU waiting handshake acknowledgement from FPGA firmware. Since FPGA firmware is broken, it never send acknowledgement (or sends incompatible handshake response) and MCU waits for infinite. That's my point of view.

Also, the update program ITSELF fails telling "Failed or not responding" immediately. I assume the update is handled by the stm32, so to me it feels like that one is buggered.

Yes, update is handled by STM32, but you're needs to enter into bootloader mode and confirm enter into bootloader on device. Without bootloader update tool will show you "failed or not responding" error or something like that.

But since STM32 is waiting response from FPGA, it cannot complete initialization and you cannot enter into bootloader mode. So, broken FPGA update can brick your signal generator.


Your PC can see USB serial port because signal generator use USB-serial chip CH341. STM32 is not involved in such detection.

Try to run ARB wave update tool without read/write option:

Code: [Select]

PSG9080_ARB COM1
where COM1 is a serial port of your PSG9080.

It requests device info from PSG9080. If it doesn't respond, then it's bricked.

The bricked device cannot be fixed with firmware update tool and requires direct firmware update through JTAG and EEPROM programmer. Unfortunately there is no RAW firmware files for PSG9080 in open access, so you cannot restore STM32 firmware.

I did get a new firmware file from our Chinese friends. This one contains an update ending in 56 78 so it should be an mcu update. But it cannot be placed due to the firmware upgrade issue.

The firmware link that you posted before:
http://68.168.132.244/PSG9080%20Firmware%20upgrade.zip

has the file name

Code: [Select]

0-256okh.bin
and it has 12 34 bytes at the end of file, this is ID for FPGA update type.

The file MD5: 8BFA094E33A656F1C2B75C206638BF7E

There is no MCU update for Juntek PSG9080 device. At least I never seen it on internet.

The only MCU update for PSG9080 that I know is MCU update for Joy-IT PSG9080. But it is not compatible with Juntek PSG9080.

It looks that you run MCU update with a wrong firmware. Can you please share firmware file which you used for update?

If this is the case, then your device is now has Joy-IT PSG9080 firmware for MCU and this is the reson why it cannot handshake with a Juntek FPGA firmware.

There is no way to restore firmware for MCU, because there is no firmware in open access and I'm afraid manufacturer will reject your request to share it, because they scary for clones from other manufacturers. But in such case you can try to write FPGA firmware from a Joy-IT PSG9080 it can help to survive device under Joy-IT brand.

Warning: In any way, backup the content of flash IC before any expreiment or manipulation to keep the way to restore it in the previous state.

There is an empty socket for a pin header on the display board; any idea whether that would be a jtag thing? Or is there any other way I could get that stm32 into boot loader mode or something similar?

Yes, this is JTAG for STM32 I tested it with DMM and the pins of this pin header are going to a JTAG pins of STM32. But I think that STM32 is protected after firmware upload on the factory. So there is no way to download firmware from the chip and only manufacturer have such file.

I tried to connect to STM32 through JTAG but it didn't respond, so this is a sign that the controller is protected from firmware download. I'm not sure for 100%, because I just leaned pins to a JTAG header for a test, so may be it fails to connect due to a bad contact. But I don't see the reason why manufacturer can leave firmware open for download, because it will allows to clone device and Chinese guys obsessed with protecting against it.

[radiolistener]

Now it looks that your seller provided you with incompatible firmware update for a Joy-IT PSG9080. But your device is Juntek PSG9080. These devices have exactly the same hardware, but different firmware. And their firmwares are not compatible with each other, so it looks like the reason why your device was bricked.

You can try to ask help from a Joy-IT company. They are more responsive, at least when I wrote them email, they sent me answer in several days.

And please let us know the results to help other users with the same issue.

[fjalvingh]

Hi radiolistener,

I was sent firmware by the shop that differs from the JOY-IT firmware, and it does end in 5678. So the hope would be that this is the MCU software for this device. It is hard to say because there is no recognisable data in the file; it looks to be either compressed or encrypted. I've added in the attachment, perhaps you might find it of use.

I tried to connect to STM32 through JTAG but it didn't respond, so this is a sign that the controller is protected from firmware download. I'm not sure for 100%, because I just leaned pins to a JTAG header for a test, so may be it fails to connect due to a bad contact. But I don't see the reason why manufacturer can leave firmware open for download, because it will allows to clone device and Chinese guys obsessed with protecting against it.

Ah, I was planning to trace those today and solder a header on it. I suppose I can still try, it cannot really hurt.

It looks that you run MCU update with a wrong firmware. Can you please share firmware file which you used for update?

Well, I'm pretty sure it was the correct one. But you might be right; it not really possible to check.

[fjalvingh]

I am wondering if anyone has any idea what this chip is in the picture, and what those connectors might be?

[radiolistener]

I was sent firmware by the shop that differs from the JOY-IT firmware, and it does end in 5678. So the hope would be that this is the MCU software for this device. It is hard to say because there is no recognisable data in the file; it looks to be either compressed or encrypted. I've added in the attachment, perhaps you might find it of use.

hm... your firmware file is really marked as MCU update v1.20.

But I never seen this firmware update on Juntek site... What seller said about it?

I think it may be correct firmware, but it may needs also FPGA update file.
Probably your current FPGA firmware is not compatible with this MCU firmware.

[fjalvingh]

But I never seen this firmware update on Juntek site... What seller said about it?

Well, the seller tells me he is in communication with Juntek about the issue, and after sending me the same (FPGA) update a few times I was able to make him understand that that update really did not work. So then he came with this one a few days after.

So, by itself I seem to have both an FPGA update file AND a MCU update file. I just do not have any way to get them where they belong..

You told me you traced the mcu jtag pins; do you happen to remember the pinout? I did trace it to the pins on the unmarked chip but I cannot make much sense of it; I see most of the pins go to pins on the chip that seem to be unrelated to jtag..

[radiolistener]

So, by itself I seem to have both an FPGA update file AND a MCU update file. I just do not have any way to get them where they belong..

can you share all firmware files provided by seller?

You told me you traced the mcu jtag pins; do you happen to remember the pinout? I did trace it to the pins on the unmarked chip but I cannot make much sense of it; I see most of the pins go to pins on the chip that seem to be unrelated to jtag..

I traced JTAG connector for a front panel board with STM32 controller (with display and buttons).

The connector on FPGA board looks like XILINX JTAG, but I'm not familiar with XILINX FPGA, I'm working with Altera/Intel chips. So I know nothing about XILINX... I think unmarked IC may be some buffer to protect FPGA pins, because they are very gentle

[fjalvingh]

can you share all firmware files provided by seller?

Well, the other one is the one widely available, with the 0-256okh.bin file.

I traced JTAG connector for a front panel board with STM32 controller (with display and buttons).

Can you tell me what you found? Because I could not really connect the pinout with the assumed stm32f103rct6. The pins I traced do not seem to go to jtag pins.

[radiolistener]

Can you tell me what you found? Because I could not really connect the pinout with the assumed stm32f103rct6. The pins I traced do not seem to go to jtag pins.

I didn't save my results, but as I remember this connector has +3.3V, GND, SWDIO, SWCLK and RESET. I trace it with DMM on STM32 pins, but I don't remember exact pin description. 3.3V and GND you can easily find. SWDIO and SWCLK can be traced with DMM.

[fjalvingh]

I didn't save my results, but as I remember this connector has +3.3V, GND, SWDIO, SWCLK and RESET. I trace it with DMM on STM32 pins, but I don't remember exact pin description. 3.3V and GND you can easily find. SWDIO and SWCLK can be traced with DMM.

Well, I could not link any of the pins to any reasonable function on the STM32. My crystal goes to pins 7 and 8 (assuming the 1-16 pin row is oriented downwards. The pins on the header go to:
1 goes to 53 (pc12)
2 goes to 52 (pc11)
3 goes to 49 (pa14) EN 48 (Vdd)
4 goes to the bottom, around pin 24, 25
5 goes to 50 (pa15)
6 goes to 51
7 is Vcc +3.3V
8 is GND

With the exception of the power lines all of this seems nonsense. But then I looked up your teardown images and lo and behold: I think I have a different version. My board looks like the image attached. The chip's markings have been erased, but clearly the pinout is quite different.

So, I fear I need to go on the hunt for what chip this is before I can do anything....

[TheBay]

That display ribbon isn't soldered on there properly  , are the connections at the top even connected to the PCB?
I guess they must have hand soldered that part in production.

I didn't save my results, but as I remember this connector has +3.3V, GND, SWDIO, SWCLK and RESET. I trace it with DMM on STM32 pins, but I don't remember exact pin description. 3.3V and GND you can easily find. SWDIO and SWCLK can be traced with DMM.

Well, I could not link any of the pins to any reasonable function on the STM32. My crystal goes to pins 7 and 8 (assuming the 1-16 pin row is oriented downwards. The pins on the header go to:
1 goes to 53 (pc12)
2 goes to 52 (pc11)
3 goes to 49 (pa14) EN 48 (Vdd)
4 goes to the bottom, around pin 24, 25
5 goes to 50 (pa15)
6 goes to 51
7 is Vcc +3.3V
8 is GND

With the exception of the power lines all of this seems nonsense. But then I looked up your teardown images and lo and behold: I think I have a different version. My board looks like the image attached. The chip's markings have been erased, but clearly the pinout is quite different.

So, I fear I need to go on the hunt for what chip this is before I can do anything....

[fjalvingh]

Ok, perhaps this rings a bell for someone?
The pin 1 marker seems to be in the top left. In that case I have found the following pin assignments:
pin 1: 3.3v
pin 16: 3.3v
pin 23, 24: XTAL/Oscillator
pin 32: ground

and of course the 8 pin pin header on top, pins numbered from 1 to 8 from left to right:
1 goes to 53 (pc12)
2 goes to 52 (pc11)
3 goes to 49 (pa14) EN 48 (Vdd)
4 goes to the bottom, around pin 24, 25
5 goes to 50 (pa15)
6 goes to 51
7 is Vcc +3.3V
8 is GND

So far I found no controller family with something resembling these connections..

[radiolistener]

Well, I could not link any of the pins to any reasonable function on the STM32. My crystal goes to pins 7 and 8 (assuming the 1-16 pin row is oriented downwards. The pins on the header go to:
1 goes to 53 (pc12)
2 goes to 52 (pc11)
3 goes to 49 (pa14) EN 48 (Vdd)
4 goes to the bottom, around pin 24, 25
5 goes to 50 (pa15)
6 goes to 51
7 is Vcc +3.3V
8 is GND

With the exception of the power lines all of this seems nonsense. But then I looked up your teardown images and lo and behold: I think I have a different version. My board looks like the image attached. The chip's markings have been erased, but clearly the pinout is quite different.

wow, your PCB is really different than my one. My one has different JTAG connector.

So there are different front panel PCB exists. And it may be the reason why firmware update was failed.

I suspect that your PCB may use different microcontroller... it may be some kind of Chinese microcontroller

[ptluis]

1) Duty cycle don't works properly for the frequency 1'000'000 Hz and above. If you set 999'999 Hz or below that it works ok. But when you set frequency just 1 Hz above, the duty cycle setting is no more works as expected. This is obviously firmware bug, probably just some kind of mistyping in the source code or something like that. But be aware, with current FPGA firmware 1.20 you will be unable to setup duty cycle for the frequency 1 MHz and above. You can setup duty cycle for higher frequency on the display UI, but the actual signal output will be invalid. It stays close to 50% for any setting from 0 to 99.99%, but it is not exactly 50%, the actual duty cycle changes for about 2-5% instead of 0-100%.

Hello radiolistener, has this problem been solved? I was thinking about buying this generator, but after reading this I suspended my purchase. I usually use square wave above 1 MHZ with variable duty cycle to test some sensors and stuff.

The PC software on the Juntek website is corrupted and has a virus. Is there another site where I can download it? I need the software to create and add other waveforms that don't exist on this device.

I'm using the FY6800 at the moment. In your experience how does the PSG9080 compares to FY6800?

Regards

[radiolistener]

Hello radiolistener, has this problem been solved? I was thinking about buying this generator, but after reading this I suspended my purchase. I usually use square wave above 1 MHZ with variable duty cycle to test some sensors and stuff.

No. It needs to update firmware in order to test if this bug was fixed in a new firmware. But I'm afraid to do it, because it may brick device and I don't have backup file to restore it. One people already reported that he bricked his device with firmware update. Therefore I am afraid to try firmware update without backup file for JTAG programmer.

The PC software on the Juntek website is corrupted and has a virus. Is there another site where I can download it? I need the software to create and add other waveforms that don't exist on this device.

Yes, the virustotal reports a virus for software from official site. But actually it doesn't needs. I tested it in isolated virtual environment (for safety purposes against viruses) and found that official software just allows to control signal generator from PC.

The only feature that is really needed is ability to download and upload arbitrary waveforms.

So, I created a small command line tool which allows to download and upload arbitrary waveform with no needs to install software. It is available on github, include source code:
https://github.com/qrp73/PSG9080_ARB

and works very well. It allows to upload and download waveform in two formats:
- 14-bit format (waveform is uploaded and downloaded from device as is, with no conversion), this is default format for my tool
- 16-bit format (it needs to convert data from 16-bit to 14-bit before upload), this format is used by official software

The conversion 16 to 14 bit which is used in official software may be unwanted if you want better control on precision of the waveform data, because 16 to 14 bit conversion is lossy and you may want to control how this loss should be done. This is why I added support for 16 bit format just for compatibility with files taken from original software. By default my tool works with 14 bit format and upload data into device as is with no conversion, so you can take full control on waveform data. But you can still use 16 bit format files to use waveform files taken from original software. It is possible by using specific command line options.

My tool doesn't have graphical editor, because I'm too lazy to do it. This is just a command line tool, it allows to download arbitrary waveform from PSG9080 into file on your PC or upload arbitrary waveform from file on your PC into PSG9080 and that's it.

The file format is pretty easy, this is just a text file with decimal unsigned values separated by new line separator. It consists of 8192 lines (8192 points) with values from 0 to 16383. You can create such file with MATLAB, R or other software.

You can use this command line tool from windows, linux or mac-os.
For linux and mac-os just install mono package and start the exe file with mono.

I'm using the FY6800 at the moment. In your experience how does the PSG9080 compares to FY6800?

I don't have FY6800, so I cannot compare. At a glance, PSG9080 has more clean and stable waveform, because I was seen worse signal from other users with FY6800.

Also PSG9080 has very nice front panel with large display and nice buttons. So, if your eyes vision is not perfect, it's better to use PSG9080, because it has more large text on display.

[ptluis]

No. It needs to update firmware in order to test if this bug was fixed in a new firmware. But I'm afraid to do it, because it may brick device and I don't have backup file to restore it. One people already reported that he bricked his device with firmware update. Therefore I am afraid to try firmware update without backup file for JTAG programmer.

That's what I was afraid of. So the seller was trying to deceive me when they said "there's no such problem". Unfortunately I can't use this device without a full working duty cycle. It's better then to find other device while still using the fy6800.  Concerning firmware updates I've read your post and the other user that bricked his unit. Both brands have the same problem on their websites, corrupted, non working firmware. This looks like feeltech, their software doesn't work anymore for firmware or software updates because they change the domain name  so no more automatic updates (not even manual). This automatic firmware upgrade is something I'm not really a fan...

Yes, the virustotal reports a virus for software from official site. But actually it doesn't needs. I tested it in isolated virtual environment (for safety purposes against viruses) and found that official software just allows to control signal generator from PC.

The only feature that is really needed is ability to download and upload arbitrary waveforms.

So, I created a small command line tool which allows to download and upload arbitrary waveform with no needs to install software. It is available on github, include source code:
https://github.com/qrp73/PSG9080_ARB

and works very well. It allows to upload and download waveform in two formats:
- 14-bit format (waveform is uploaded and downloaded from device as is, with no conversion), this is default format for my tool
- 16-bit format (it needs to convert data from 16-bit to 14-bit before upload), this format is used by official software

The conversion 16 to 14 bit which is used in official software may be unwanted if you want better control on precision of the waveform data, because 16 to 14 bit conversion is lossy and you may want to control how this loss should be done. This is why I added support for 16 bit format just for compatibility with files taken from original software. By default my tool works with 14 bit format and upload data into device as is with no conversion, so you can take full control on waveform data. But you can still use 16 bit format files to use waveform files taken from original software. It is possible by using specific command line options.

My tool doesn't have graphical editor, because I'm too lazy to do it. This is just a command line tool, it allows to download arbitrary waveform from PSG9080 into file on your PC or upload arbitrary waveform from file on your PC into PSG9080 and that's it.

The file format is pretty easy, this is just a text file with decimal unsigned values separated by new line separator. It consists of 8192 lines (8192 points) with values from 0 to 16383. You can create such file with MATLAB, R or other software.

You can use this command line tool from windows, linux or mac-os.
For linux and mac-os just install mono package and start the exe file with mono.

I manage to get the full working software from tony albus youtube page. I also download your software but can't test it because I didn't purchase this generator  but tell me how do you draw the waveforms?

I don't have FY6800, so I cannot compare. At a glance, PSG9080 has more clean and stable waveform, because I was seen worse signal from other users with FY6800.

Also PSG9080 has very nice front panel with large display and nice buttons. So, if your eyes vision is not perfect, it's better to use PSG9080, because it has more large text on display.

Correct PSG9080 it's better than fy6800 cleaner waveform from what I saw in video reviews, apart from jitter, amplitude stability and wave distortion was solved by using a linear power supply made from a old transformer and some 7812, 7912 and 7805 linear regulators.


I find out who manufactures PG9080, Hangzhou Junce Instruments Co., and did you know they are releasing a new brand name for their FG? ready? JunCtek    This chinese guys don't realise what words can do to a product 

[radiolistener]

That's what I was afraid of. So the seller was trying to deceive me when they said "there's no such problem".

The seller may be right. I got my device about year ago, so a new device may come with the more fresh firmware and it is possible that this bug is already fixed. I don't know.

but tell me how do you draw the waveforms?

I don't paint it, because it is not precise method. I just write my own code to generate it from a function or script. The same you can use MATLAB, R, python or other language to do that. You can also plot the waveform from a file in such way.

[ptluis]

I don't paint it, because it is not precise method. I just write my own code to generate it from a function or script. The same you can use MATLAB, R, python or other language to do that. You can also plot the waveform from a file in such way.

thank you for the info 

[Neper]

Correct PSG9080 it's better than fy6800 cleaner waveform from what I saw in video reviews, apart from jitter, amplitude stability and wave distortion was solved by using a linear power supply made from a old transformer and some 7812, 7912 and 7805 linear regulators.

I'm running my 9080 with an external linear PSU connected to the 5 V DC input. Its built-in PSU injects far too much noise back into the mains.

I find out who manufactures PG9080, Hangzhou Junce Instruments Co., and did you know they are releasing a new brand name for their FG? ready? JunCtek    This chinese guys don't realise what words can do to a product 

My favourite is still Rigol. Supposedly from the French 'rigoler' or 'rigolo'.

[radiolistener]

I'm running my 9080 with an external linear PSU connected to the 5 V DC input. Its built-in PSU injects far too much noise back into the mains.

I power it from a USB 10000 mAh power bank. But it consume high current, about 0.9-1.5 Amps, it depends on selected mode and selected amplitude, so it needs good thick power cable. With cheap Chinese thin cables voltage drop is too high and you can see that display is dimmed.

Interesting, does someone tried to install bipolar linear power supply for output driver to disable it's onboard DC/DC converters?

[ledtester]

FWIW, a video by Jerry Walker comparing the harmonics and noise floor of a couple of different signal generators including the PSG9080, a SDG1020, TTi  TGR1040 and Rhode and Schwarz SMIQ 02B.

Juntek PSG9080 Harmonics Comparison -- Jerry Walker
https://youtu.be/Vq6sbQ_OCAs

[fjalvingh]

Just for posterity: I finally managed to resurrect my bricked Juntek PSG9080, as follows:

* Switch off AC power at the back
* Keep the 8 on the keypad pressed then switch on AC power, keep 8 presse
* Keep 8 pressed and press the ON button
* Release the 8.

This should now show a bootloader on the screen. Connect to USB, press 1, and update with the files attached to this message and the Juntek upgrade tool. You need to run both files: one is for the MCU, the other for the FPGA.

Thanks a lot to this post for the method to access the bootloader > https://www.eevblog.com/forum/testgear/joy-it-psg-9080-firmware-failure/msg4559212/#msg4559212

[torch]

In case  it's of interest to readers: 

There's a Koolertron CJDS98 which appears to be the same as this Juntek 9080.

If you download the zip file software and docs package pointed to by the Koolertron vendor on amazon.com, it contains a manual titled PSG9080, and showing a picture of the Juntek.

Of course "same" is relative, as there could well be hardware and firmware variations.

There is a separate, Koolertron branded, manual. It's hosted on the same IP address as the Juntek manual. The manuals are virtually identical inside.

[Quarlo Klobrigney]

The first time I used the generator, the power supply splattered the entire MW and SW bands (don't even ask about LW) on a radio I was tying to align.


I bought it to align the big stuff (you Europeans know the type) from the 50's and 60's as it was smaller than my FY6900 that's wedged in between other pieces of equipment on the overhead bench.


That power supply is absolute trash, How do they justify selling the generator for such a price with that thing powering it.
I set it aside to put a linear supply to it at a later date. 2 years later it still awaits.....


Quote from: Neper on 2021-10-16, 10:27:49

>Quote from: ptluis on 2021-10-14, 06:38:23

Correct PSG9080 it's better than fy6800 cleaner waveform from what I saw in video reviews, apart from jitter, amplitude stability and wave distortion was solved by using a linear power supply made from a old transformer and some 7812, 7912 and 7805 linear regulators.

I'm running my 9080 with an external linear PSU connected to the 5 V DC input. Its built-in PSU injects far too much noise back into the mains.

[kosmos]

what do you think is the Joy-it PSG9080 or the JUNTEK part newer or preferable. I ordered the Joy-IT PSG9080, maybe I'll get a RIGOL DG10xxZ for comparison.

[Bravo]

I found that Norton antivirus corrupted the PC software. When I used a laptop which does not have Norton on it there was no problem.

[Edison]

The source in the PSG9080 doesn't just make a mess of the network and the output, I hear quite high tones and I can't sit near it when it's on, I normally hear a whistle from the transformer - for now, operation at 5V - I plan (two years) to convert it to a linear source and finish the input for EXT 50MHz

Just an addition: the power supply beeps audibly only without load, yesterday I thought of trying to see what the small plate can withstand - the power supply shut down at 4.3A and supplied 3A without a significant increase in temperature

[radiolistener]

yes, mains PSU module used in this signal generator has issues. It has audible noise and may inject some noise interference into mains. I recommend to replace it with something better quality PSU module or just power it from external 5V / 2A PSU. You can power it from USB power bank.

[Muxr]

I don't have a sig-gen that can produce such high frequency (other than my RF signal generators). Was interested but part numbers being lasered off is a big turn off for me. How far we've come.

We went from quality equipment and companies which would provide beautiful service manuals with schematics and even provide you with a roll of factory solder to companies lasering off, off the shelf component part numbers. Such a shame.

[Edison]

module or just power it from external 5V / 2A PSU. You can power it from USB power bank.

Attention! the minimum requirement for a DC power supply is 3A, during normal operation it can easily reach 1.8A under load, and at the moment when another command comes, it will easily exceed the limit of two Amps and distortion may occur due to insufficient power supply. As I already wrote, the original source can handle 4A without a problem.

Nice day 🙂 Tom

[Edison]

So I did a little research, an audio quality Ultra Low Noise linear power supply will easily swing over $100 (think with professional grade mechanical modifications). It is a question of whether this investment is worth it and whether an ordinary linear source with slightly better filtering will suffice.

First I'll do another measurement to see how much noise from the source affects the generator board, from what I've found so far the main board itself is producing some noise.
Thanks to the oscillator's fine correction function, the effect of the external reference is not satisfactory as Tony Albus found out while editing

[Edison]

I almost forgot: I have the GOUPCHN logo on the device, but inside it says Juntek - HW says 1.2 and the board says 1.3

[RUS]

Hi All

Has anyone tested sweep frequency on VCO mode ( to connect sawtooth to MOD input  5v p/p offset 2.5 v)?  It works?

[radiolistener]

So I did a little research, an audio quality Ultra Low Noise linear power supply will easily swing over $100 (think with professional grade mechanical modifications). It is a question of whether this investment is worth it and whether an ordinary linear source with slightly better filtering will suffice.

I think there is no sense to use 100 USD power supply here. You can use cheap but good quality power supply.

Note that there is DC/DC step-up converter to get bipolar power for output amplifiers. If you want to get the best quality signal, there is a sense to disable onboard DC/DC and replace it with external bipolar linear PSU. I believe such mod can give you the best signal that can be achieved on such kind of type signal generator.Also it should fix clipping issue which happens sometimes at 25 Vpk output with 50Ω load.

[Edison]

Note that there is DC/DC step-up converter to get bipolar power for output amplifiers.

The switching power supply replacement project was suspended for this reason. The DC/DC converter will likely be retained for the 5V external power option, but the power supply will be reworked to be linear when powered from the mains.