Designing a reverb driver on drawing table - what sounds good?

[überfuzz]

I'm trying to fix a reverb circuit I did a while back. This time I'm trying a simulation approach, i.e calculating cut frequencies, setting up a spice simulation etc. I've got quite good insight in how to control the signal in the reverb circuit. But no clue of what a normal reverb frequency range is. Nor how much reverb signal should be mixed in, or better, what the ratio normal/reverb signal should be.

I understand that it's in many ways a matter of taste, but just to get into the ballpark. I tried to google reverb frequency range and such, but with no real luck.

[jazbo8]

Have a look at this.

[R.G.]

There's a section on reverb tank drive and drivers with frequency responses in the old National Semiconductor Audio applications book. It's oriented towards using semiconductor power amp chips for drivers, but the drive level and frequency response tidbits are not semiconductor specific.

One issue they make is that you need to drive lots of current into the input to get the signal level up. That would be the one item that I'd worry about with the no-transformer approach. True that you can do it without a transformer, but you might want to raise the current drive level to get better S/N.

[J M Fahey]

Also googled, after 10 pages of useless images had to give up very disappointed.

Really, Internet is sh*t and I mean it.

Zillions of tons of useless superficial information, but no real deep knowledge available.

Or so deeply buried that amounts to the same.

I remember having seen spring frequency response graphs, how come none shows up when googling them?

Anyway, just from memory, they have thousands of peaks and dips, the less flat response I have ever seen, much much worse than any speaker (which are already spikey).

If you are comitted to it, build a small spring driver amp, say an LM386 or a TDA2030 (much better), a flat 10X or 50X gain recovery single Op Amp recovery stage, the approppriate 1/8" plug cables and sweep it with your PC soundcard.
REW, Audacity or a host of speaker tracing programs provide the signal and tracing.

One caveat, you need an actual continuous tone sweep, say a sinewave which goes slooooowwwwlyyyyy from, say, 50Hz to 10kHz in at least 10 t0 20 seconds and measures what comes out at the other end.

Many modern ones work with a single pulse or chirp which lasts a fraction of a second and do a lot of number crunching later to calculate the frequency response.

Pretty useless here because of the time delays involved (the gate might open and close, before the pulse travels end to end) and remaining resonances, measured in seconds, also play a big part.

That will show the voltage driven curve, then if you wish you can repeat it with a current driven one.

Then post them here

Another MEF "first".

Personally I want to do that test , problem is that I have another million cool tests in the queue (like most in MEF), so not sure when it will actually run ... you might help

[Leo_Gnardo]

Along the lines of what Juan said ^^^, you have reverb tanks that vary one to the next, even in the same batch. I wouldn't sweat over the drive circuit, especially as a way to panel-beat lousy sounding springs into good ones. You'd be better off using a standard drive circuit. Buy tanks by the dozen and listen to each one, select the best for your amps and discard the rest, or sell them off to someone who doesn't care.

[überfuzz]

Maybe I was unclear. I'm not aiming for a flat reverb, rather, what frequency range should be there, i.e. what bandwidth should the drive circuit have. My comment on the flat driver circuit was a comment on the jazbo8 linked.

At the moment I'm trying to figure out what current level I have to get to saturate the reverb tank I building around. Knowing this I can design the stage of the reverb driver that is going to feed the signal to the reverb tank.

[jazbo8]

Maybe I was unclear. I'm not aiming for a flat reverb, rather, what frequency range should be there,

5kHz is often mentioned.

[Tooboob]

Many reverb tank solutions aren't driven hard enough. Drive it hard and get a fast decay tank unless you're into surf music.

[überfuzz]

Everyone, thanks for the input!

How do I get info about the signal needed? Like if I wanna know voltage*current.

[Jazz P Bass]

Some helpful information here: ::::::::: Accu Bell Sound Inc :::::::::

Go to the header that says 'Application'.

[Mick Bailey]

I've experimented quie a bit with reverb and concur with the need to get plenty of drive and CC. Scoping the signal into the coil can tell you quite a bit when you compare it with the impedance curve. It's quite difficult to get this just right so that there's enough drive at higher frequencies. As for the drive frequency range, I use a 470pf cap on the input to my 1st stage with a 470k grid resistor. Too much low-end sounds really muddy. Think of how natural reverb sounds - there's pretty much no bass.

An area often overlooked is also limiting the bottom-end at the recovery stage; if you have full-bandwidth recovery the reverb becomes over-sensitive to handling noise, footsteps and booming, plus it will feed back uncontrollably at higher volume. Trimming the recovery-tube coupling cap to 1nf or less transforms the reverb and gets rid of the LF spring noise.

My own mix contol goes from 100% dry to 100% wet. It seems excessive, but I can get the theme to the film 'Get Carter' easily.

The Accutronics site has good info on driving reverb trays.


Edit: Ah - that's the link Jazz P Bass posted.

[Chuck H]

Some helpful information here: ::::::::: Accu Bell Sound Inc :::::::::

Go to the header that says 'Application'.

Nice to see they put some tech info back on the site. It seemed to be missing for awhile before the face lift (and apparent name change).

[Chuck H]

According to a couple of scholarly articles I've seen from respected members here the input to a spring pan should have a LF -3dB around 400Hz. When allowed, lower frequencies would occupy the majority of current and create what sounds pretty much like atonal mush. It's also said that frequencies into the pan above about 1kHz can cause shrill, tinny reverb. I suppose it makes sense to try and reverberate as much fundamental as possible since reverb has a mass of phase error produced harmonics anyway. I do ok rolling off a bit higher but the concept still applies. So the challenge comes from generating enough current at what is approaching mid range frequencies. Which might explain why so many reverb drivers look to be too powerful for the tiny transducers in the pans!?! Like a transformer driven 6K6, for example.

I honestly haven't built that many reverbs, but a couple have been projects that required great consideration on the subject you're asking about. There are many variables in the speakers, cabinets, how the actual drive and recovery circuits balance for the final reverb level, possible transformer limitation. In my own experience there will be some EQ adjustments to your final design for tone and/or stability no matter how carefully you simulate and plan the design. If you look at a diverse array of reverb designs it's easy to see that there are many which fly in the face of ideal parameters and sound fabulous anyway! So idealize some if it makes you happy. I did and I'm not sorry for it. If your experience ends up like mine you may find that some idealized limitation needs to be undone to some degree to make it sound "right" instead of bland.

[MarkusBass]

If you search for: "spring reverb driver frequency response" in google, you will find several good articles on this topic. In most cases they list high pass filter frequency at least 200 Hz, or 300Hz and low pass filter frequency 6 kHz, but sometimes also 1 kHz.
Have a look:
Spring Reverb
The Valve Wizard
https://www.amplifiedparts.com/tech_...d_and_compared
UA WebZine "Ask the Doctors!" April 06 | What Makes Spring Reverbs Sound "Springy"?

Mark

[überfuzz]

Thanks for all the input. I'm still not sure of how hard I should drive the driver stage. For that I need the current needed to dive my tank. However, I read in one of the links that over-saturating the input won't add any significant distortion. If that's true I guess I can focus on the pick-up end of the circuit, and the frequency response.

Mick Bailey - If I get your numbers right you have a high pass filter with a cut off frequency of 720Hz. Seems high, but if it sounds good I guess.