Quick reverb driver theory question

[Ch0jin]

Going on from a previous thread about weak reverb in a couple of Peavey amps I decided to school myself on all the intricacies of solid state reverb drive/recovery circuits, and I have a question.

This is the "standard" reference design (from accutronics) for a general purpose driver using a push pull amplifier after the op amp. (cheaper circuits drive the tank direct from the op-amp)

http://members.iinet.net.au/~ch0jin/stuff/accu_verb.jpg

What is gained by using reverse parallel diodes for biasing, and adding the 4k7 resistor R56 as per the commonly used Peavey version below?

http://members.iinet.net.au/~ch0jin/stuff/reverb.jpg

[Ch0jin]

OK after some reading I may have an idea as to part of the answer to my question.

In the Peavey schematic CR9 and CR11 are without doubt being used to bias Q3 and Q4, and when combined with R57 and R58 as temperature compensation, form a textbook Class AB amplifier optimised to eliminate crossover distortion.

What you don't see in 99% of class AB amplifier examples are CR8 and CR10. In my amp these are 1N4148's by the way. The posted schematic is quite old as it uses germanium diodes.

Anyway, I was looking through lecture notes online and I found a similar, but not identical implementation as flyback protection for the opamp.

The example I saw was a theoretical explanation and had the two extra diodes connected directly from the output of the op amp (pin 7) to Vcc (pin 8) and Vee (pin 4) rather than through the current limiting resistors R54 and R55, but it's the closest thing I've seen to an explanation.

The theory being, that attempting to make a sudden change in current flow in an inductive load, (such as a reverb tank) will cause large voltage flyback spikes. These flyback spikes appearing on the output of the op amp can destroy the output stage of the amplifier.

If I'm right, it goes some way to reinforcing the widely accepted notion that Peavey make bullet proof gear, as they are the only manufacturer I've found so far using this.

I'm still lost on R56 though. Maybe some kind of bias feedback arrangement?

So this leaves me with the understanding that of the possible implementations of reverb drive circuits, the Peavey version is probably the most rugged and potentially powerful designs, but I still don't know conclusively why the reverb depth and character are so weak compared to other examples I have heard. Next stop, the tank itself and the recovery circuit.

[jim p]

Here is a good posting on reverb http://sound.westhost.com/articles/reverb.htm that you may want to check out. To drive a reverb tank correctly it is best to be current driven which is how the tank is driven if the reverb tank input is in the feedback loop of an op amp. With reverb tanks other than the 8 ohm input impedance tank using an op amp to drive the tank is no real problem. I think all the Accutronics information is from the 60’s when op amps did not have the same current capability as they do now. Plus if you use a dual op amp you can run two of them in parallel to drive the input transducer on the reverb tank. Notice on the Accutronics schematic they have an RC in the feedback of the op amp to increase the voltage gain with frequency to make it act more like a current amplifier then a voltage amplifier. You should look at the recovery circuit as well as the input circuit you want a gain of approx 100 on the recovery side.
I have used the op amp feedback to drive tanks with no problem, I have also built a hybrid voltage amplifier like you have pictured, also a transistor driven reverb transformer circuit they all work fine. Have also implemented the Belton digital reverb in an amplifier which sounds good. The input to the reverb tank only requires milliwatts of power so unless you are going to drive an 8 ohm tank a NE5532 should have plenty of current capability to do the job. Fender has all there schematics on their web site to see how they have implemented reverb in there amps.

[jim p]

I failed to notice before but the Accutronics circuit is a voltage amplifier while the second circuit is a current driver. The feedback on the Accutronics driver is connected at the emitters of the two buffer transistors. While on the second circuit it is connected across a 10 ohm current sense resistor R72, so as the impedance of the reverb input rises the voltage at the output will rise while the voltage across R72 will remain constant. So on the second circuit you do not need to cause the gain of the amplifier to rise with increasing frequency

[Ch0jin]

Thanks Jim, I actually have read over a bit of that page already and bookmarked it for further study. It's been a long time since I studied this stuff in college :) I need to read up on voltage amps V current amps as that is a common topic in reverb design it seems. That's a good tip on the feedback circuit of the Accutronics design. I knew it made the feedback gain frequency sensitive, but wasn't really sure why. Now I get it.

I have actually been looking at recovery circuits too and have several here to study, I was using Peavey schematics because I have one (well two actually but one doesn't have reverb) I'll definitely check out some fender schem's too. I've been studying ancient fender schematics for a while as I'm toying with the idea of gutting my home made JCM800 and rebuilding it properly as something more closely aligned to it's true heritage (I built it from the burnt out head of a 1960's Aussie made VASE bassman, so I'm thinking as the original circuit is long gone, a 5F6a would be a suitable homage)

Do you have any thoughts on those extra diodes? Could I be on the right track with flyback protection? Also R56 the 4K7, does that have something to do with the voltage amp Vs current amp configuration? It just seems out of place for a Class AB config to me.

Thanks again :)


Also....my first post on the new forum :)

[jim p]

I was thinking of the flyback also but that should be taken care of by one of the two transistors it would just forward bias then on harder. I suppose they can’t hurt and they may have been a response to having failures occurring. I do like the idea of the current drive it is how a pentode would drive the tank through a transformer. You would set the current relative to the input signal by the value of R72. As I said before if you use say a 190 ohm input impedance tank putting the reverb tank in the feedback loop of an op amp works just fine. The down side of a higher input impedance tank is you run out of voltage compliance at a lower frequency but for guitar 3 kHz to 4 kHz should be plenty. The most recent circuit I did for reverb is here http://www.sewatt.com/node/20778 over at Sewatt. Triodes actually make poor drivers for reverb tanks because of the Early effect causing there gain to change with output signal.

[Ch0jin]

Thanks again Jim. I'll go and check out that circuit also.

[jim p]

If this is a tube amp on the simple side I have posted this over at Sewatt http://www.sewatt.com/node/20854 Depending on the location of the volume control and signal levels changes might need to be made in the basic circuit.

The big problem I see with reverb is the tank manufactures have never put much time in specification and application information. They provide you enough to get going then you need to make corrections to your assumptions if required. I think they could sell more reverb tanks if they did a better job of explaining the electrical operation and requirements. Also some application circuits that are newer then ones done in the 60’s or 70’s might help also.