UNI-T UTG932 Signal Generator Vpp issues

[Marlos]

Dear EEVBlog community,

this is probably a beginner question, but I recently bought the UNI-T UTG 932 signal generator and was trying to test it with a scope (Hanmatek DOS1102).
 
I am trying to check this recently bought, cheap-ish gear by generating simple sinus or square waves with the siggen in the freq. range 1 kHz - 1 MHz.
I set the Vpp of that wave to around 2V-4V (offset 0), and connect the signal generator output to one of my oscilloscope channels (normal probe at 10x, probing the BNC-to-crocodile-jack output that came with the siggen, i.e. NOT via direct BNC-BNC coax cable).
When I do this I don't have that typical 2x Vpp on the scope when you forget to set the siggen to HighZ instead of 50ohm, but more than 10x attenuation accross the whole range I'm testing. That is, I am reading around 90mV - 150 mV on the scope for 2-4Vpp setting on the siggen, and it doesn't get much higher than that even if I set the Vpp on the siggen to 10V.

Yes, siggen is set to HighZ output, and both the scope probe and the scope setting are 10x attenuation. 
I know the scope is OK from testing the self-calibration output (5Vpp / 1kHz square wave). Could my signal generator be faulty or am I missing something obvious?
Edit: frequency reading from siggen into scope is fine but I can't for the life of me get that Vpp and offset reading to match.
I also used the 121GW meter to check the Vrms with open siggen output. Same as the scope reading: wrong compared to what I see on the siggen display.

Thanks for any help

[Mechatrommer]

check cable continuity. probe directly on siggen's BNC output without cable. show picture if possible so we can see clearly your setup. is siggen output light is green?

[Marlos]

Continuity is good, probing directly at Siggen BNC output gives same results.
Here's a couple pics from the smartphone.

1st one is the siggen with the croc jacks being probed by my scope's CH1 (too big for 2 pics).
2nd one to be posted soon: scope (CH2 on self-calibration outputs). I put Vpp and f measurements on screen, should be visible.

[Marlos]

Have been having issues trying to upload this second pic, gonna try it ZIPPED this time:

[Mechatrommer]

so it seems you have a broken siggen. you can either (1) return the unit to the seller and ask for replacement or refund, or (2) open the unit and see if anything wrong. your problem with attachment is you have too big pictures (4608 x 2304 pixels), try reducing picture size in smartphone to make smaller picture and easily uploaded to internet or forum.

[HB9EVI]

seems odd, that's true; if you have it from a local supplier and still under warranty, send it back; otherwise try to fix it. if you're not too far away, I could maybe help you out...

[Marlos]

Hi folks,

thanks for the help so far.
Fact is, I bought this off AliExpress. I got in touch with the seller to see if they play ball (I even asked for a service manual  .
If they don't answer by the weekend, I'll try to fix it myself. Any tips on what to check first once I start taking this thing apart?

Best,

[HB9EVI]

ok, tough luck I guess; even if you get the seller to agree with some kind of warranty, you'll carry the costs of the shipping back to him, plus all the pain with VAT and customs.
I never even bothered trying to get warranty in such cases but directly got me a screwdriver.

you have to trace the signal path starting from the bnc socket backward

[Mechatrommer]

attached is the UTG962 pcb (i suspect same as yours). you can start by checking around the output opamp THS3092, the relays, BNC port connection to pcb, the 50 ohm output resistor etc... btw, if sending back for full refund is not an option, try repairing it, if success, you can blackmail the seller for partial refund or risking dispute with aliexpress "buyer protection" management team. or better, simply forgive him it maybe not his fault. usually i will just ask politely for partial refund or simply give bad feedback as a warning to future buyers. ymmv.

[Marlos]

Thanks fellas.
I'll give it a go soon as I can, then come back here to let you know.

At my fellow Swiss resident: thanks for the offer to help out, but I think I'd like to try to repair this myself and probably just grab a proper quality siggen for 50 bucks more, if I can't.
Tough to be riding between here and Jura with the current Covid wave anyway

Funnily enough, I got a PSU from AliExpress recently and it came "broken" as well. I realized it wasn't broken by checking the outputs, but rather the LED screen was dead.
When I opened it up: a loose ribbon cable, and nothing more 

[HB9EVI]

sure, I meant more online by mail or so; however I guess you should be able to fix that.
I stopped getting things from ali; the recent 1-2 years more and more articles got lost while being shipped; I stuck only with shops where I had no problems so far.
test equipment I don't buy there anymore at all

[Marlos]

Hello fellas,

indeed this is a tricky one for a noob like me. Lots of scrubbed off components and also I'm slightly unaware of how signals should look like exactly at each point.
Only ever used siggens that were working, first time poking into one

However! I am measuring 5 V and 0 V at what I believe should be an output of + / - 12 V at the places where Mechatrommer indicated in his pics of the PCB.
I am probing ground at the wire that's floating between jumpers above and through the FPGA heat sink. I.e. the ground from the 5V/2A power supply.

I imagine that the +/- 12V are with respect to that ground, please correct if wrong?
Pic attached: note there is 0.5V DC offset and Vpp is very weak (frequency is correct but Vpp should be 4V).
This is reading either channel of the siggen on CH1 of the scope
CH2 of the scope showing 5V at +13 V probe point on the PCB.

Further probing reveals:
5V where it says +13V
0.75V where it says -13V
5V further back where it says 5V
0.75V further back where it says -5V.
1.2V and 3.3V probe points are OK.

Looks to me like those negative power rails are the issue, since having a DC offset and too low Vpp happens identically on both channels. Am I on the right track?
Some solder joints look like they can use a reflow but continuity is still fine wherever I found this.

Turned power off. Forward-bias probing the diodes with the 121GW gives me 0.18 V on both, reverse-bias probing them gives 1.5V / 2.1 V.
In resistance mode, forward-bias on both diodes is giving me 2 kOhm. Reverse-bias is 100-250 kOhm on the "upper" diode (Mechatromm's pic frame of reference).
The "lower" diode is showing 2M-ohm reverse-bias resistance.
Not sure if those values on the diodes are as bad as it looks to me, or if it's normal because of the particular +/- power rail circuit configuration.
I did try to uncharge all caps before probing the diodes (power off ofc).

The last thing I tried was checking that XL6008. It's supposed to be the voltage regulator right?
Enable is high (5V), feedback pin 0.5-0.75V, both Vin and Switch out are a constant 5V. It looks to me like switch out should definitely not be a constant 5V?
I changed the Vpp to 10V on the siggen, but after hearing relays click I'm reading the exact same thing out of the voltage reg.
Could this be the culprit or at least clearly one of them?

Any further tips on how to find out exactly which component is failing how?

Also from my fellow Swiss resident, I am having trouble finding shops that easily deliver electronics components to private customers.
Stuck with Conrad at the moment using work e-mail. Any other tips? Is there a cool secret Swiss radio shack I don't know about?

I feel close to a solution here. Or is it more like  ?
Either way sorry for the delay, life gets in the way sometimes. Thanks a lot for the help. And a great 2021!

[Mechatrommer]

The last thing I tried was checking that XL6008. It's supposed to be the voltage regulator right?

XL6008 is smps controller in boost mode. you should check its output first after confirming it has 5V on its input. its output should give square wave (switching pulse)

Further probing reveals:
5V where it says +13V
0.75V where it says -13V
5V further back where it says 5V

from my experience, this indicates either the smps controller (XL6008) is not switching due to damaged chip or overloaded on its output. first check resistance of +13V and -13V to ground to find if any short. if there is no short, maybe XL6008 is damaged. but replacing XL6008 is the last thing you should do after you ruled out anything else problem in the circuit. since smps controller will got damaged if its overloaded anyway. and sometime smps will recover if you remove any bad things on its output, i dont think XL6008 is one of the cheap crap smps knockoff IC, but well, who knows whats the XLSEMI is. good luck.

Turned power off. Forward-bias probing the diodes with the 121GW gives me 0.18 V on both, reverse-bias probing them gives 1.5V / 2.1 V.
In resistance mode, forward-bias on both diodes is giving me 2 kOhm. Reverse-bias is 100-250 kOhm on the "upper" diode (Mechatromm's pic frame of reference).
The "lower" diode is showing 2M-ohm reverse-bias resistance.

diodes are also important aspect in smps of that type. not just they have to work properly, they have to be low drop type, ie schottky type. you should really check all diodes are working properly, unsolder and check them individually if needed. replacing them is cheap, but buying stock locally can be problem, need to buy from online store.

[Marlos]

And again with great delay, some news.
I de-soldered diodes and checked, they are all good. Soldered them back in, same problem.
Soldered them back out and soldered brand new ones in - exact same behavior afterwards.

As I said in a previous post about the SMPS:
Enable is high (5V), feedback pin 0.5-0.75V, both Vin and Switch out are a constant 5V. It looks to me like switch out should definitely not be a constant 5V? -> this is still the case.

Also I am attaching some pictures here illustrating the following (please note that the output voltage is never in any case the one displayed on the siggen):

• From 1V up to 5V, the waveform is capped at the bottom with a "wavy" pattern.
• At 5V I hear a relay click and then the waveform looks right but still the peak-to-peak is nonsense.
• After 8V the top of the wave is the part that gets a weird wavy pattern (different from the one in the first case).
• This doesn't go away anymore however high I place the output voltage setting (and the distortion gets worse and worse).

Can it be that SMPS after all? Can it be something much simpler that my untrained noob eyes aren't seeing?

As usual I appreciate all the input - including recommendations of better siggens, since by now the time I spent on this would justify buying a new one. Although it was a nice experience and pushed me to get some SMD gear that I was too lazy to get before to do the basic troubleshooting.

Something that is available to me would be the Joy-it JDS6600 but if my fellow Swiss resident can recommend something better than Conrad from whence I can get gear like that for decent prices, I am all ears and thankfulness.

Peace! And thanks for the support!

[lachko973]

Hello! Just registered to share my experience with the exactly the problem.
After measurements it came up that the voltages +13 -13 and -5 are missing.
When the XL6008 PWM controller is heated up with soldering iron or hot air gun the voltages are coming back and the generator works perfectly.

Couldn't find XL6008 around  me so I replaced it with XL6009.

Best regards !

[Yuriy V.]

Couldn't find XL6008 around  me so I replaced it with XL6009.

How did you do it when 6008 is much smaller than 6009?