AD831 Chinese Mixer Module Questions And Answers

[Quarlo Klobrigney]

Before I hook this up, can anyone tell me the specs of these modules from sources such as Aliexpress and eBay https://www.ebay.com/itm/0-1-500MHz-High-Frequency-AD831-RF-Mixer-Drive-Amplifier-Module-Radio-Converter/153996768961 ?
I see really no info except for the raw IC in the AD datasheet: https://www.analog.com/media/en/technical-documentation/data-sheets/AD831.pdf
It specifies a LO and RF input without max input specs.
A supply voltage of: Dual supply ±4.5V-±5.5V or Single supply 9V-11V

So if anybody has used this module do you have any recommended parameters?

EXTRA: From a CERN paper page 44, 5.2.3 AD831 active mixer: "optimal LO drive level is -10 dBm. However, increasing the LO level to 0 dBm results in lowering the unwanted spectral components" Who knows, it's a paper by the people who want to suck us into a black hole...

[Quarlo Klobrigney]

I still, after searching all day for anybody using this module and found that all of the suppliers have comments like "Works great! just what I was looking for" and "This really helped me on my design" and the rest....

One guy on the tube says 5V, another one on there 9V and so on, No definitive answers on what voltage or max input levels to apply. Although their projects work, they are from other countries and don't speak English and offer no wiring diagram help....

Anybody in the other hemisphere awake and reading this can help, I'd appreciate your help. I have to tell the vendor soon that it either works or get money back for a defective product. Thanks.

[vk4ffab]

The datasheet is pretty clear. What specificaly dont you undrstand.i will answer your questions. I have the modules here and have used them

https://robs-blog.net/2020/03/02/ad831-mixer-module/

[_Wim_]

From the PCB picture I see that pin 4 and 5 are tied to GND, so VN (negative voltage) is at GND. This means single supply operation. The pcb  "GND" and "Vcc" are directly connected to the AD831, so no local regulation or resistor divider. This means you have to power the board between 9V and 11V according to the datasheet. 10V is thus recommended.

Looking at the pcb, it seems they used pretty much the "single supply operation" schematic of the application note section of the datasheet (see attachment). The only differences I can spot are the input filtering, which is a single capacitor instead of C-L-C, and no low pas filtering is installed (82pF caps are omitted).

Using the attached schematic you can easily trace out if anything else is different, and when you are sure you have the complete schematic, read the appropriate sections of the AD831 datasheet to fully understand how you can use this. As said above, the AD831 datasheet should be quite clear. 

[Quarlo Klobrigney]

Well I did power another module ordered in April initially at 5V and only one port would pass (forgot it it was the LO or RF). It was well below the datasheet max dBm level if you can believe the datasheet as it contradicts itself as to level input specs in a few places.
This is on a well regulated supply. I noticed that the AD831 IC was very warm, to which I found out that more than many are concerned about in other non answered postings, including an AD China site.
It then quit outputting anything. I then tried a 10V supply using a 7810 (I had a few) with good regulation and filtering but it still would not output anything.
The input was a signal generator knocked down by a pi attenuator to a proper a level on a scope somewhere around 60MHz.
I don't want to be accused of blowing this module up by the vendor. Vendors such as "weflowerbotique", "manythanksyou", who haven't a clue as to what they are selling.
I bought it instead of constructing a diode ring and associated torroids. I have used SBL-1's before and that was a very nice passive solution.
So now you can understand my apprehension to repeating failure.
If I could see one of these things working, it would alleviate such fear.

[vk4ffab]

Well I did power another module ordered in April initially at 5V and only one port would pass (forgot it it was the LO or RF). It was well below the datasheet max dBm level if you can believe the datasheet as it contradicts itself as to level input specs in a few places.
This is on a well regulated supply. I noticed that the AD831 IC was very warm, to which I found out that more than many are concerned about in other non answered postings, including an AD China site.
It then quit outputting anything. I then tried a 10V supply using a 7810 (I had a few) with good regulation and filtering but it still would not output anything.
The input was a signal generator knocked down by a pi attenuator to a proper a level on a scope somewhere around 60MHz.
I don't want to be accused of blowing this module up by the vendor. Vendors such as "weflowerbotique", "manythanksyou", who haven't a clue as to what they are selling.
I bought it instead of constructing a diode ring and associated torroids. I have used SBL-1's before and that was a very nice passive solution.
So now you can understand my apprehension to repeating failure.
If I could see one of these things working, it would alleviate such fear.

With my tests with it, i ran it at 5v with a -7dbm 1Mhz LO input and -80dbm 7mhz RF input. It worked as designed with the + and - signals on the IF port at 6 and 8mhz.

[FenTiger]

I have one of these modules too, though I haven't done anything with it yet.

The data sheet says the quiescent current is 100 mA. With a 10V supply that's 1W of power. It looks like it's expected to get a bit warm!

[Quarlo Klobrigney]

I also forgot to add that the first vendor stipulated that I send a video of the module not working to YouTube, which I don't have. Even if I did, it is an easy thing to fake by cutting a wire and sliding the PVC jacket over the cut (I wouldn't do such a thing) and show what? Show a scope, SMA cables to and fro and a power supply? How does one capture all that? So I never got a refund. Not the first time burned by Aliexpress. One time was with leaded 1+ MΩ 1 % resistors 1/8W sized that were out by 10 & 20 %. I had to send a photo of every value (10 or so various) with my Fluke LCD showing the discrepancy. At that distance 5 band colours cannot be seen. I vowed never again. A colossal waste of energy and time and money. So they keep the money I get 0.

Anyway, I don't know that I will hook it up again any time soon due to the pucker factor. Perhaps I will await some more nice people here (not that I don't appreciate everyone so far) to chime in with more test results, video, blog, etc. before I bite the bullet. I'm not young and daring, I'm old and cautious.

[vk4ffab]

Done just now. The IC should not be hot, it should be at ambient temperature. I could do you a video but I really do not see how that would make any difference. Everything you need to know is contained in those images and the datasheet. So the module has been running for the last 15mins and the surface temperature of the IC is 21 Deg C, a few degrees above ambient temperature. The in and out impedance's are 1.2K I have not tried to match them, but doing so would improve the performance of the mixer. Sorry the images are blurry, stupid phone and arthritic hands.



Power == 5V



RF == 0.0001v LO == 0.1v




Sum and difference frequencies.



The module under test.


GOD I HATE HOW THIS FORUM DEALS WITH IMAGES

[Quarlo Klobrigney]

Wow! You have gone the extra Mile/KM! You see, what you have done is what I couldn't do because my Feeltech was never sent after ordering back in April, 2 times now!
BTW the 3rd one is sitting in an Amazon parcel locker as we speak for pickup in the morning. Dishonest or downright nasty sellers who steal your money for months and deliver nothing. Well that's in another post. I paid the extra $$ and it's here in less than a week.

The signal sources I were using were 2 old Wavetek 148 20MHz and 166 50Mhz function generators, that were unstable, bulky and old. I then went to a combination of function generators and oscillator modules. I knocked the oscillator modules down with pi attenuators to the right levels of course.
If I only had what needed from the beginning, it certainly would have been as easy as you appear to make it.
I appreciate your photos and setup. I don't have my HP spectrum analyser any more as I sold it for living expenses a few years back. Things are better now. Hopefully in the future a Rigol or other will fall into my lap.
You have done a most excellent job in explaining the process without anything blowing up.
I thank you. You have given me the confidence to duplicate this experiment tomorrow if the Feeltech FY6900 #3 finally exists and functions properly. Stay tuned to this frequency....
Now, it's time for bed.

[Quarlo Klobrigney]

Well the original module does not work. The new module at 5VDC is outputting a heterodyned signal as viewed on an oscilloscope. I don't have a spectrum analyser as I have said before. It will vary proportionally to a change in ether LO or RF frequency as well as LO signal level changes. I have tried anything from 2-50 MHz. So the next step is to fire up an SDR on a laptop and look at it there. The problem is workbench real estate to do all of this. Lots of other projects going on at the same time with immediate deadlines. This includes workbenches 1-7. I will report back here with findings soon.
I managed to dig out a hole on the bench and do the tests with my finally arrived FY6900 order #3. I "bolted" the module and 7805 down to prevent "accidents". Crude, but effective.

[radiolistener]

I still, after searching all day for anybody using this module and found that all of the suppliers have comments like "Works great! just what I was looking for" and "This really helped me on my design" and the rest....

I have this module, it works
Bought it about two years ago on aliexpress. Works as expected.
I found that it works good as a receiver with about -30...-40 dBm LO on the LO input.

[radiolistener]

One guy on the tube says 5V, another one on there 9V and so on, No definitive answers on what voltage or max input levels to apply.

You can find specs in the datasheet: https://www.analog.com/media/en/technical-documentation/data-sheets/AD831.pdf

[Quarlo Klobrigney]

Yes, the datasheet.... It says 9V-11V or depending on the external circuit ±5V.
The question is which configuration is this green module and the other black module? 5V would be safe but what if it can be operated at 10V? Some listings say the 9V-11V in the description.
Also, I was injecting about -35dBm into the LO port, and -47dBm into the RF.
When I get some time I will reverse engineer a schematic from the board I have.
 

[Quarlo Klobrigney]

I can't abide by unsolved mysteries. So I drew the module up in Eagle. The capacitors that don't have values are ones that I cannot measure. Anything marked NU is not installed. The schematic basically follows the AD datasheet for a 9VDC input voltage.
Here is the AD schematic and my redraw of the green module from China:
(It won't let me post 2 inline expandable pix for some reason!)

[Quarlo Klobrigney]

BTW, I have yet to try it on 9VDC. So here the other one is:

[radiolistener]

BTW, I have yet to try it on 9VDC. So here the other one is:

Are you sure that L1 is installed on your module? On my module it is missing. There is a place for soldering, but no element...

C5 = 90 nF

Also, on my module pin 2 and pin 3 connected together but missing capacitor C7. There is a place for soldering C7, but no element. The same thing with pin 19 and 20, they connected together, but missing capacitor C8. There is a place for soldering C7, but no element.

Also it looks like your schematic has mistake. C7 and C8 should be connected to +9V. On your schematic it is connected to the ground...

[Bud]

According to the datasheet the left side of R3 in your schematic should connect to pin 16 of the IC, not pin 17.

[_Wim_]

I think you are also missing an output capacitor (Cc) in your drawing. From what I can see in the picture you posted, it seem to be installed.

[Quarlo Klobrigney]

I drew it up quickly as I had to get some things done. It was Saturday after all...
I misspoke when I said CX & RX I meant anything that said NU = not used. The board does have a place for it but it not have it installed. This like any schematic should represent this missing component regardless. Eagle does not offer reference numbers to have the same name, i.e. Cx & Cx is illegal. I did correct the CX & RX text description to NU. I then added the missing output cap in the schematic which was missing. Any reference numbers in the Eagle schematic have no bearing to any reference numbers on the board.

[Quarlo Klobrigney]

You are correct they go to +9VDC. I did have them in the proper orientation but I think that's when the phone call came in and I inadvertently connected them to ground.

Quote from: radiolistener on Today at 13:20:27>Also it looks like your schematic has mistake. C7 and C8 should be connected to +9V. On your schematic it is connected to the ground...

[Quarlo Klobrigney]

I did try the module on a faulty (low 8.56V) 7809 regulator and it is working. the only issue is hum - ripple on the signal. Hmm I wonder where that could be coming from FEELTECH! So I'm going to have to void my warranty and do some fixes in the FY-6900. Then hook up the SDR Rx and have a look at the spectrum.

A quote from the AD datasheet: Single-Supply Operation
Figure 10 is similar to the dual-supply circuit in Figure 9.
Supplies may be as low as 9 V but should not be higher than 11 V, due to power dissipation.

[Quarlo Klobrigney]

Well here's the story so far after reading the scattered data in the datasheet. The typical power supply voltage and min-max levels as well as the low pass filter calcs are in the attached photo. I made an LibraOffice Calc sheet.
Somebody ready to try it? I already have, for function of output only. I'm busy with some rush repairs so I can't hook up the SDR to see the harmonics and mixer products yet.

[hpw]

Hmm,

currently dealing with this modul and may you share

. latest circuit

. xls sheet

while my project is somewhat different: 2 oscillators (low PN to come) and mixing down to 200Hz ... 40kHz range.
So the maximal Vpp would be an issue too, while have about 4Vpp

Hp

[vahegan]

Quarlo has not been posting in this thread for quite some time, and his spreadsheet is not available. However, the datasheet tells us how to calculate the lowpass filter capacitors.


So, it is very easy to calculate the filter capacitors: Cf=5684/f
where f is the output (difference) frequency in MHz and resulting Cf will be in picofarads