Help me diagnose an op amp reverb driver circuit I built for belton "digilog" module

[elipsey]

Hi,

I built a solid-state driver circuit for a Belton "digilog" reverb module and installed it in my home-brew amp. I don't really know what I'm doing with op-amps, but I tried my best to Follow Instructions. The pre-amp is built in the fashion of a Marshal, and the power amp is p-p 6V6s, with LTPI. Except for the messed up reverb, I think it sounds pretty good. The reverb driver circuit includes a 5VDC supply derived from the filament supply using a bridge rectifier and 7805 linear regulator, as well as an lm358 dual op-amp dip used to buffer the send and return signals.


Power supply:



Driver circuit, installed in the amp:



A 512K reverb mixing resistor is installed after the first gain stages and before the cathode follower pair, and the reverb send and return are connected before and after this mixing resistor. The Belton reverb module seems to be working correctly; it's input and output pins look clean, and reverb-y, respectively, on my scope. However, the op-amp that buffers the wet return signal seems to be outputting a half-wave of it's input. There is gross distortion in the signal after the reverb mixer, which I believe results from the half-wave reverb return signal being superimposed on the original pre-amp signal at the mixing resistor. Of course, maybe there is something else going on that I don't understand... In any case, when I disconnect the reverb return, the signal after the mixer looks better, and the amp sounds normal again.

Schematic from Belton docs


Here's a link the Belton design documentation for the digilog, which I got this schematic from:
https://www.smallbearelec.com/Projects/BTDR-1H.pdf R1 and C1 are 330k/5000pf for a 200hz high pass. R2 is 47k, expecting 6-8VAC at the input, which seems to be in the neighborhood of what it actually gets. R3 is omitted. The 5V supply measures 4.9VDC which is well withing spec for the reverb module, seems like it should be fine from the op-amps, and it's pretty much free of ripple. I tried disconnecting the resistor that connects the high-pass filter input to the non-inverting (+) input of the second op-amp (see schematic), but it didn't seem to help. Does this just mix the dry signal back into the output, or do something subtler like NFB, or something? I don't really understand why it's there.

I double checked that layout of my little DIP circuit board pretty carefully for mistakes, solder bridges, &c, although I suppose there could still be a mistake there that I missed.

Can anyone suggest what I might try next?

Thanks in advance,
Chris

P.S.
This question has already been posted on TDPRI, but I'm not getting much traction there; I don't think most of those guys are into this kind of stuff. I hope it's not rude to link to the images I uploaded to the other forum...

The input looks ok



Half-wave output at return, why?



Wet output looks nice and wiggley, but still half-wave

[loudthud]

That circuit needs plus and minus supplies on the opamps.

[elipsey]

That circuit needs plus and minus supplies on the opamps.

thanks
i have +5 on Vcc, so like -5V connected to the lm358 GND pin, but the signal ground and all the other stuff is the same? and that would let the output swing negative? i notice that +/- 12V seems common for op amp audio amplifiers in various schematics.

the manufacturer's sample implementation schematic doesn't show anything about the power supply for the op amps... i assumed the GND pin in the pinout meant 0V

err... what's a convenient way to get -5V? maybe i have to start over....

[loudthud]

If you are trying to shoehorn this thing into a tube amp, that is a really bad idea. Sorry I have to tell you that. Opamps do not work well in tube amps without lots of protection diodes. They do not like the big signal voltages. The LM358 won't sound very good no matter what you do.

Your best course of action is to add the reverb circuit between the guitar input jack and the first tube. Use a TL072 and +/- 15V (or +/-12V) rails. The 7805 will work fine to drop +15V down to +5V for the Belton module. Does the power transformer in the amp have a bias tap?

[teemuk]

How is it powered, single-ended or dual supply? Is proper bias provided for the opamps?

There seems to be no AC coupling at all. This may not be a good idea because every gain stage will amplify DC offsets of their inputs. As gain multiplies per each stage even tiny offsets in input may become huge DC offsets in output.

[elipsey]

If you are trying to shoehorn this thing into a tube amp, that is a really bad idea. Sorry I have to tell you that. Opamps do not work well in tube amps without lots of protection diodes. They do not like the big signal voltages. The LM358 won't sound very good no matter what you do.

Thank you for replying. Yes, that is exactly what I'm trying to do. I thought I might learn something.

When you talk about protection diodes do you mean that they would be arranged to clip the input signal at (maybe at .7V) to protect the op amps? Why don't op amps like big signal voltages? What's big? Will they brake or just clip?

Your best course of action is to add the reverb circuit between the guitar input jack and the first tube.

Problem: the first gain stage is not always the same triode. This amp is set up like a JCM800 (if you have ever looked at one of those), where there are two gain stages before the cathode follower, and plugging into the low input jack just skips the first one. I put a master volume before the PI to make the levels reasonable. So the signal voltage can vary by an order of magnitude or so at the reverb mixer. I knew that would be a problem, but I thought it would be fun to try the reverb anyway. I was thinking of removing one of the jacks and adding a relay for channel switching such that the signal is attenuated in high gain mode; that might level out the two channels somewhat. It's an experiment.

Use a TL072 and +/- 15V (or +/-12V) rails. The 7805 will work fine to drop +15V down to +5V for the Belton module. Does the power transformer in the amp have a bias tap?

The power transformer does have a bias tap, it's about 37VAC I think. So should I connect that to a voltage divider and then use a pair of rectifier diodes (and filters) to get +/- 15V? Except for the whole thing being a bad idea, am I on the right track here?

Why is the TL072 better than the LM358?

[elipsey]

Thank you. I guess it's a single ended supply, if that means that no negative voltage is provided? I have +5V on the Vcc pin, and 0V on the GND pin of the LM358. I used the schematic provided in the Belton sample implementation and, to the best of my ability, did not deviate from said schematic. So.... no AC coupling at all. Now that you mention it I can see why that could be bad.

I don't know how to bias an op amp. I was wondering if that was the function of the 10k resistor between GND and the (+) input of the op amp on the right hand side of the schematic.

So why don't I try putting a coupling cap before the input of the whole thing, and using values and an input voltage similar to what was in the documentation? Surely they didn't provide a reference circuit that would never work for anyone ever. If I did that the only unknown, with respect to duplicating the reference circuit, would be the op amp type, and power supply. What unstated assumptions might the authors have made about those things?

[g1]

The documentation shows the supply voltage for the Belton module only. It is assumed you will be using another split supply of whatever is required for the type of op amp you are using.
There is no AC coupling (or input biasing) shown for the op amps because it is assuming op amps in use with split supply, not single ended.
They do mention "replaces spring unit" so I think they are also assuming there is an existing op amp reverb circuit.
Yes, getting a split supply from the bias tap would be a good way to go, assuming it is not being used to provide bias for the amp.

[Dave H]

I don't think it's a good idea to power it from a bridge rectifier across the filament supply. The filament supply will have its centre tap grounded which means the negative output of the bridge can't be grounded.

[g1]

I didn't see where anyone mentioned the filament supply? Using the extra bias tap (if equipped and unused) was mentioned.

[Dave H]

The OP mentioned it in his first post g-one.

[52 Bill]

And with the opamp power supply being so low, driving it with a 6-8 volt input signal will be another problem.

[nickb]

I really don't think it is possible to give you a complete answer without knowing more. A schematic of your home brew would be big help. For sure there's a bit more to it that you relalised

Do you need to switch it in and out?
How much room do you have on your chassis?

My inclination is to use a dual rail to rail low noise opamp like TLC2272 on the input fed off a single 5V rail connect in the guitar input. Some redesign of the example circuit will be required.

We can derive a 2.5V mid-point level to drive the reverb chip. Looking at the sample cct I would say the inputs and outputs of the reverb chip are AC coupled so we don't need to worry too much about DC levels on it's i/o's.

I don't think using the filament supply is a good idea because of possible mid point ground issues but also I've seen noise introduced into the filament by the rectifiers cause problems that are hard to solve. The 37V tap if regulated to 5V could work but the reverb chip uses lots of current and will result in power dissipation/heat problems.

If there's room a new small 5VA heater transformer with 7805 type regulator giving a single 5V supply could be the least troublesome route.

To try it out to see if you like it you could use four 1.5V batteries as a temporary supply and feed that through a diode to drop it below the 5.5V limit on the Belton chip.

Anyway, tell us more about it & post a schemo else we're just guessing .

[elipsey]

First of all, thanks you guys are awesome

The documentation shows the supply voltage for the Belton module only. It is assumed you will be using another split supply of whatever is required for the type of op amp you are using.
There is no AC coupling (or input biasing) shown for the op amps because it is assuming op amps in use with split supply, not single ended.
They do mention "replaces spring unit" so I think they are also assuming there is an existing op amp reverb circuit.

Ok thanks, I guess that explains it.

Yes, getting a split supply from the bias tap would be a good way to go, assuming it is not being used to provide bias for the amp.

Nope, it's not being used.

I don't think it's a good idea to power it from a bridge rectifier across the filament supply. The filament supply will have its centre tap grounded which means the negative output of the bridge can't be grounded.

Well, the filament supply has a virtual center tap, so the center tap (or centre tap, for those across the pond) is not connected. If it's still a bad idea to use a grounded bridge rectifier on the filament supply, then I should stop, because I'm doing it now!

I really don't think it is possible to give you a complete answer without knowing more. A schematic of your home brew would be big help.

Ok, fair enough. I will get a schematic together and post it, post-haste.

For sure there's a bit more to it that you relalised

Yes, I like it when there's a bit more to things than I realize.

Do you need to switch it in and out?
How much room do you have on your chassis?

No, a level knob will do, and some.

Here's a picture of the whole chassis:


The node that oscope probe is connected to is the return side of the 512k reverb mix resistor, which is also the top half of a voltage divider, formed by two 512k resistors that resides between the first two gain stages and the cathode follower. The red wire at that node is the reverb return, which is connected to the output of the second op-amp in the Belton sample schematic. On the upper left, you can just see the Belton digilog module, and the little upright board on the lower right holds my rectified 5V supply, and op amps.

My inclination is to use a dual rail to rail low noise opamp like TLC2272 on the input fed off a single 5V rail connect in the guitar input. Some redesign of the example circuit will be required.

We can derive a 2.5V mid-point level to drive the reverb chip. Looking at the sample cct I would say the inputs and outputs of the reverb chip are AC coupled so we don't need to worry too much about DC levels on it's i/o's.

I don't think using the filament supply is a good idea because of possible mid point ground issues but also I've seen noise introduced into the filament by the rectifiers cause problems that are hard to solve. The 37V tap if regulated to 5V could work but the reverb chip uses lots of current and will result in power dissipation/heat problems.

If there's room a new small 5VA heater transformer with 7805 type regulator giving a single 5V supply could be the least troublesome route.

To try it out to see if you like it you could use four 1.5V batteries as a temporary supply and feed that through a diode to drop it below the 5.5V limit on the Belton chip.

Anyway, tell us more about it & post a schemo else we're just guessing .

I have another little DIP board like the first one that I'm sure I could squeeze in somewhere to hold a dual-ended power supply for the op amps. I'd rather not add a transformer if I can find another way to do this. I like the idea of deriving a +/- 15 from the bias supply, and using the 7805 to get +5 from that. I suppose I can wait patiently for new op amps in the mail if necessary. I'm still fuzzy about why one sort of op amp is better than the next.

If anyone would like to direct me to an accessible primer on solid state/analog audio or op amp design patterns, I would be grateful.

I'll post a schematic as soon as I can. Thanks again, you guys kick ass.

[loudthud]

This power supply should get the job done. The +/- 15V is good for about 15mA, the +5 more than 100mA.

The LM358 is a very slow opamp not suitable for audio of any quality. The TL072 is more than 10 times faster. It's used in lots of solid state guitar amps.

The circuit on the Belton data sheet mixes the dry and wet signals. It is intended to be inserted inline with the signal chain.