Why LEDs create higher sound volume compared with other diodes when used in Ring modulator

Nigel Goodwin said:

The load shown in your circuit of post #3 is 10K, the load should be whatever you wanted the output impedance of the transformer to be. Lowering the load resistance also lowers the input impedance of the transformer. A load resistor should also flatten your peak out somewhat, and potentially might improve your bandwidth as well?.

Click to expand...

with the 10K load attached the increasing voltage problem is gone. The upper -3db point is 20Khz so it mostly covers the whole hearing spectrum. I did not measure lower -3db point since it was below 10Hz. Is this normal with this many windings or did I somehow got lucky.

I can't really hear over 18Khz but I think my ears are still in decent shape since it is very high lol.

Fluffyboii said:

while Germanium diodes have around 0.7V voltage drop.

Click to expand...

Silicon diodes have about a 0.7V drop.
Germanium diodes have only about a 0.3V drop.

rjenkinsgb said:

Audio transformers for telephone modems are usually 1:1 and pretty good at audio as the modems needed quite wide bandwidth with good frequency response and linearity.

Click to expand...

The "wide" bandwidth is only 4 KHz !!!

Fluffyboii said:

with the 10K load attached the increasing voltage problem is gone. The upper -3db point is 20Khz so it mostly covers the whole hearing spectrum. I did not measure lower -3db point since it was below 10Hz. Is this normal with this many windings or did I somehow got lucky.

I can't really hear over 18Khz but I think my ears are still in decent shape since it is very high lol.

Click to expand...

I would expect the higher number of turns to give an extended low frequency response, beyond what is needed.

Ramussons said:

The "wide" bandwidth is only 4 KHz !!!

Click to expand...

The audio sample rate is just 8KHz, so the speech bandwidth is 4KHz or less - however V92 modems use synchronous data transmission with multiple levels on each sample at 8KHz, so the bandwidth and linearity of the coupling transformers must well above that.

From the V92 standard:

6.2 Symbol rate
The upstream symbol rate shall be 8000 symbol/s derived from the digital network.

Click to expand...

To reach the maximum 56000 bits per second the modem pair must accurately encode and recover seven bits per symbol, so any of 128 levels in each 8KHz sample or symbol...

crutschow said:

Silicon diodes have about a 0.7V drop.
Germanium diodes have only about a 0.3V drop.

Click to expand...

1n34 diode compared to a silicon small signal one give similar Vd
Datasheet says forward voltage is max 1V. Should I buy another diode. 1n270 datasheet also says max 1V.

Are you particularly wanting a germanium diode for something? - and where did you buy the one you have?. I suspect it was yet another fake from China?.

Nigel Goodwin said:

Are you particularly wanting a germanium diode for something? - and where did you buy the one you have?. I suspect it was yet another fake from China?.

Click to expand...

I bought it from a legit store. It was quite expensive. I can see the inside of it being probably a vacuum with a very small wire in the middle. People originally used these for ring modulators so I wanted to try it as well. I can also get some schottky diodes like bat85. It is something I need to test.

Nigel Goodwin said:

Are you particularly wanting a germanium diode for something? - and where did you buy the one you have?. I suspect it was yet another fake from China?.

Click to expand...

Today I went to the store to test all diodes with my tester device. It turns out many Germanium diodes actually have voltage drop more than 0.2V. Most of them were similar with silicon diodes around 0.6-0.7. Then they gave me a new bag of them with different model number which is AAZ15 instead of 1n34 and they were around 0.35V so I picked the closest ones with my cheap component tester. I also got some 1N5819 schottky diodes which have Vd around 0.25V so even better then Germanium ones. I am sure they are not selling fake products. I wonder why Germanium diodes differ in voltage drop this much.
I wanna have some of those AA117 diodes 0.18Vd lol. Unfortunately it is hard enough to find a shop that sells Germanium diodes.

I am using this circuit without the CNULL and SNULL pots and I get extreme carrier leaking to output when I use diodes other than LEDs. LED don't cause this problem and it is silent as it should be when there is no input sound unlike other diodes. Why I get this problem with Germanium and Schottky diodes but not LEDs? All diodes are matched carefully.
Also modulation seems like not happening with other diodes and all I hear is distorted voice with carrier leak. That means It will probably not be any good if I add CNULL pot ant try to negate carrier leak.
I used tl074 and tl072 for this circuit. Not that it matters just saying.

That circuit looks wrong to me...

The carrier drive amp feedback is taken from the wrong place and it cannot regulate diode current; it will go from nothing to clipping over a tiny range, no linearity or proportionality to the input signal.

Try connecting the carrier opamp output direct to the left transformer centre tap and put the 330R resistor in series with the ground of the right transformer, then pick up the carrier level control feedback from the right centre tap.

That makes the feedback actual current, and should allow the carrier gain to directly control the diode current, with it be proportional to the input signal, all the way down to near zero. It should make it far nearer linear.

A carrier null pot is also pretty essential with that type of circuit.

Make sure the "unbalance" pot it at 0V.

rjenkinsgb said:

That circuit looks wrong to me...

The carrier drive amp feedback is taken from the wrong place and it cannot regulate diode current; it will go from nothing to clipping over a tiny range, no linearity or proportionality to the input signal.

Try connecting the carrier opamp output direct to the left transformer centre tap and put the 330R resistor in series with the ground of the right transformer, then pick up the carrier level control feedback from the right centre tap.

That makes the feedback actual current, and should allow the carrier gain to directly control the diode current, with it be proportional to the input signal, all the way down to near zero. It should make it far nearer linear.

A carrier null pot is also pretty essential with that type of circuit.

Make sure the "unbalance" pot it at 0V.

Click to expand...

Are there any other changes needed to be done in order to cope with the 750 ohm 3400 turn transformers I got as a mistake. I actually used 150 ohm resistors instead of 330 because I thought they were for preventing the op amp from shorting and it wouldn't matter much.
"pick up the carrier level control feedback from the right centre tap." how I do this?

Fluffyboii said:

"pick up the carrier level control feedback from the right centre tap." how I do this?

Click to expand...

The right hand end of the "CL" pot would connect to the centre tap of the right-hand transformer, rather than the left-hand one, so it is sensing the voltage across the [moved] 330 Ohm resistor than now goes from the RH centre tap to ground.

The rest should be OK, I think; it's trial and error..

rjenkinsgb said:

Try connecting the carrier opamp output direct to the left transformer centre tap and put the 330R resistor in series with the ground of the right transformer, then pick up the carrier level control feedback from the right centre tap.

Click to expand...

Is it something like this? Maybe other diodes than LEDs are not working because diodes get saturated as you said and they stay open all the time.

rjenkinsgb said:

The right hand end of the "CL" pot would connect to the centre tap of the right-hand transformer, rather than the left-hand one, so it is sensing the voltage across the [moved] 330 Ohm resistor than now goes from the RH centre tap to ground.

Click to expand...

Are these messages contradicting each other? Should the carrier op amp output go to left or right middle tap and if it goes to right what left middle tap should go, is it like the second one?
It is hard to draw with phone.

Sorry, right centre tap.
Like this:

rjenkinsgb said:

Sorry, right centre tap.
Like this:

View attachment 138414

Click to expand...

I will try this as fast as I can and tell you about the results. BTW how did you changed the schematic like that, it looks original. I took that from a wiki page so I thought it was accurate. I wonder why they go with inverting aplifiers and waste another op amp for inverting it back.

I used photoshop, drawing and erasing lines & copy/pasted the resistor and ground symbols.

I have never built one of those but I have designed an awful lot of gear over the years, and that part just seems wrong to me, as it was trying to control voltage linearly which meant the diodes would never conduct at low carrier levels, messing up the linearity.

Sensing the current through the transformer winding and diode combination should mean it controls current in proportion to the carrier level, so hypothetically a more linear input.

Whether it works better or not is another thing! I'll wait for you to try it out.


I have come up with a totally different & possibly original way creating a ring modulator, hopefully as good as a an AD633 but with only simple low-cost parts. I've not had chance to build one to try it yet, too many other projects!
(The key part is one of these, but not used as an amplifier:
https://www.ebay.co.uk/itm/154964259020 )


Re. the double opamps: The first TL072 on the output is a unity gain inverter.

It allows the "null-ing" signals from the two null pots to be mixed to the input of that and the input of the next stage, so by moving each null pot end to end both the level and the phase of the signal added in can be adjusted, to find the point that gives best carrier or source cancellation.

rjenkinsgb said:

I used photoshop, drawing and erasing lines & copy/pasted the resistor and ground symbols.

I have never built one of those but I have designed an awful lot of gear over the years, and that part just seems wrong to me, as it was trying to control voltage linearly which meant the diodes would never conduct at low carrier levels, messing up the linearity.

Sensing the current through the transformer winding and diode combination should mean it controls current in proportion to the carrier level, so hypothetically a more linear input.

Whether it works better or not is another thing! I'll wait for you to try it out.


I have come up with a totally different & possibly original way creating a ring modulator, hopefully as good as a an AD633 but with only simple low-cost parts. I've not had chance to build one to try it yet, too many other projects!
(The key part is one of these, but not used as an amplifier:
https://www.ebay.co.uk/itm/154964259020 )


Re. the double opamps: The first TL072 on the output is a unity gain inverter.

It allows the "null-ing" signals from the two null pots to be mixed to the input of that and the input of the next stage, so by moving each null pot end to end both the level and the phase of the signal added in can be adjusted, to find the point that gives best carrier or source cancellation.

Click to expand...

Well I bought a AD633 too. I think it is fake since it starts heating up for no reason and I got it for 3usd but I may try it later. It is really expensive here around 20usd. I would love to know another way of making a ring modulator tbh just because it is very fun to use.

rjenkinsgb said:

I used photoshop, drawing and erasing lines & copy/pasted the resistor and ground symbols.

I have never built one of those but I have designed an awful lot of gear over the years, and that part just seems wrong to me, as it was trying to control voltage linearly which meant the diodes would never conduct at low carrier levels, messing up the linearity.

Sensing the current through the transformer winding and diode combination should mean it controls current in proportion to the carrier level, so hypothetically a more linear input.

Whether it works better or not is another thing! I'll wait for you to try it out.


I have come up with a totally different & possibly original way creating a ring modulator, hopefully as good as a an AD633 but with only simple low-cost parts. I've not had chance to build one to try it yet, too many other projects!
(The key part is one of these, but not used as an amplifier:
https://www.ebay.co.uk/itm/154964259020 )


Re. the double opamps: The first TL072 on the output is a unity gain inverter.

It allows the "null-ing" signals from the two null pots to be mixed to the input of that and the input of the next stage, so by moving each null pot end to end both the level and the phase of the signal added in can be adjusted, to find the point that gives best carrier or source cancellation.

Click to expand...

I applied the changes. It made the carrier leak worse and introduced signal leak as well. And I can hear carrier leak with LEDs too. But with the other diodes there is some modulation which wasn't there before I think. Now I will install nulling pots and see if I can get it to perform well. btw I realised when I was doing changes that I connected signal level pot directly to op amp output instead of across 330 resistor that was connected to op amp out. Maybe that was why it didn't worked well at first.

Nulling pots made it better but there is still some considerable carrier leak that I can't null with pots.

I should try reverting it back but before that I will wait for your reply.

The diodes I am testing all have 230-231mV voltage drop. My test device cant measure difference of 1mV or lower accurately.

Also LEDs still produce sound that is at least a magnitude higher in volume than other diodes which was my main question in this topic. It is interesting.

The reason why other diode didn't work at first time was because the carrier gain was to high and carrier pot should been almost at minimum side. Leds did not had this issue since they had more voltage drop. Now I reverted the changes and it still got the carrier leaking issue. I am so stupid it was working very nice at first.

Last edit: I went back and fort with your changes and original. First of all I found that the 330 resistance doesn't matter if it is there or not. It doesn't effect even a little bit for all of them. Your changes makes all diodes easier to get modulating but it increases leak. Original carrier pot is more fiddly to get going for small Vd diodes but it got less leak. My final answer is that I played on this too much and tried fixing something that did not need fixing and now I will always have leak no matter what I do. This is my punishment for being greedy, it is not first and it will not be last. I will probably scrap this and throw all in garbage after my obsession advances.

Try swapping the connections to the ends of one of the transformer windings? The leakage is possibly from inter-winding capacitance, or otherwise something to do with the overall circuit layout, grounding etc.