Ruby Preamp for Sonic Impact T Amp

[Carl]

I want to use a Ruby as a preamp for a Sonic Impact T Amp. Can it be directly connected or will I need some kind of a buffer?

Any help will be appreciated.

[darwindeathcat]

You should be able to connect them directly together. Since the "volume pot" of the ruby comes before the IC, you may need to add a "master volume" pot (100k log would be okay) between the ouput of the ruby and the input of the Sonic T.
  You should also check to see what the gain is set to on the Sonic-T. If it is set too high, you may get some distortion when driving it with a pre. A gain of, say 20ish, would be okay, where as gain higher than 30 may begin to produce some distortion when running the pre at full volume... Remember that with solid state amps, if you want distortion, you want it in the preamp stage, NOT in the power amp... The ruby should give you this, or build you could build a noisy cricket for the pre if you want more control over the distortion than is offered by the ruby...

[armstrom]

I'm doing something similar. I will be using a modified ruby circuit as a preamp to a TDA2005 15W power amp. I will make the following modifications to the ruby circuit:
-Replace the 10K volume pot with a big muff tone control
-Add a "master volume" control on the output as the other poster recommended.

My power amp board has a 10k trimmer that acts as a sort of master volume so I may just replace that with a 50K pot as the master control rather than have two pots, one right after the other.

Here's a schematic of the power amp (this was drawn by our own N9VOC and posted over on the diyAudio forums.. I hope he doesn't mind my re-hosting it) http://www.mr2-power.com/downloads/fk607.pdf

As you can see in the schematic the VR1 10K trimpot would act like a master volume control. Do you think just replacing the trimpot with a 50K log taper will be good enough to act as a master volume? (remember, there will be no volume control between the JFET buffer and the input of the LM386)

I do have one question that maybe someone can answer... can I reduce the size of the output cap on the ruby? The 220uF seems pretty big for just coupling to a fairly high impedance power amp stage.

-Matt

[darwindeathcat]

Armstrong, looking at the schem you linked to, I note the following. The stock 10k trimmer works with the 4k7 resistor to set the input impedance of the amp, and thus work as a VC. Replacing the trimmer with a 50k log pot will allow you to dynamically change the input impedance over a larger range, but not that much larger. To be honest, you may want to bump up to a 100k pot for even better control.
The power amp schematic also has a 0.1uf input coupling cap, so you actually do not really need the output coupling cap in the ruby at all (the DC voltages will already be isolated by the poweramp's input cap). The couplng cap value matters only in that it effects the lower end of the frequency spectrum let through to the next stage. Larger cap values let more bass through than will lower values. 0.1 uf will let PLENTY of bass through, so you can go smaller (0.047 or 0.022uf) if you want less bass. Also, FYI, the VC pot you are planning to put in will work with any output cap on the ruby to create a highpass filter, so the if you choose to keep an output cap, be aware that the tone will probably change a bit as you adjust the volume.

[armstrom]

One thing I wanted to add is a preamp output jack so the amp can be plugged in to a PA for bigger sound. To do that I think I'll need to keep the 10K trimmer and add a new volume pot between the 220uF output cap of the ruby and the input of the power amp. I will probably reduce the size of the output cap though, maybe a .1uF as you suggest.

Do you see an issue with leaving the 10K trimmer after the 100K master volume control? I'm assuming enough signal will still make it to the power amp.

I can post a schematic of what I'm talking about later today.
-Matt

[darwindeathcat]

I'm actually right in the middle of doing the same thing with my LM3886 amp (but using the proffessor tweed circuit from runoffgroove.com for the pre), and was just researching the best way to add a preamp out jack. What I am going to do is to actually just have the leads for the out jack connect directly to the lugs of the master volume pot. This can be done two ways: one way has you connect to the "out" lug so that the connection comes before the VC, and the other way is to connect the jack to the "in" lug, so that the position of the VC pot has no affect on the volume of the preamp out. Here is the interesting thing, however: Provided that you build a simple bypass switch to the internal preamp, you can use the preamp out jack in reverse (that is, you plug an external preamp into it, and feed signal in to the power amp). Why do it this way? Well, 3pdt switches are expensive, and you would need to use one if you wanted a true bypass (one that shorts both signal and ground) so you could then just use the normal "in" jack in bypassed mode. But i am poor! I have some DPDT switches (much less expesive and common), and I have some stereo jacks. It is well known that when you use a stereo jack with a mono plug, you can use it as a switch (connecting the ground with ring two of the jack to make a ground connection, and thus switching on the circuit). Combine this ablilty with the simple DPDT switch and you've got a true bypass when the input jack is unplugged and the DPDT is switched over. But you need to feed signal to the power amp, so you feed it through the jack that would normally be used to send the preamp signal out to the PA... Get it?
Anyway, that's a bit complicated... To address your question, it does not really matter if you keep the 10k trimmer or not. Just know that any resistance in series adds up to the total resistance your signal will have to cross before it enters the amp. The both the VC pot and the 10k trimmer will act as variable voltage dividers sending varying amounts of the signal to ground, depending upon how you adjust them. You don;t have to totally remove the 10ktrimmer, though, you can just turn it all the way open so that it sends all the signal through...
Actually, I took a look through the ruby/little gem schematics again, and I noticed that all of them have that unusually large output cap (220uf) originally, I misread it for 220nf, which is a normal size for an output cap. I also noticed that it is an electrolitic, so it must have to do with a requirement of the lm386 rather than normal DC decoupling. In that case I would keep it there... 220uf is quite high, so I would defintately keep the 0.1uf cap in the input of your poweramp. The ruby is meant to drive speakers directly, so that might also have something to do with it.
I also noticed that the "little gem" has it's VC AFTER the lm386 http://www.runoffgroove.com/littlegem.html You could probably just do your master VC just the way it's done in that schematic. I once built a "noisy cricket" which is a ruby-type amps that has similar tone controls to what you propose to do, so I would go over to beavis audio research and read up on how he does it...
Hope that helps...

[teemuk]

About the preamp output jack: Why not just implement a full FX loop arrangement there instead. Add one more jack that serves as the preamp out and have the other one serve as the power amp in. Use a switching jack for the latter. Wire the two jacks together in a way that the jack's internal switch disconnects the signal path from the preamp out jack whenever you insert a plug to the power amp in jack. Now the jack will work as either power amp in jack (for external preamps) that disconnects the internal preamp or simply as a return jack for the FX loop. You can also use the preamp out jack for signal takeoff without disconnecting the PA. Pick a random schematic of any amp that has a FX loop and it likely shows what I'm talking about.

Put the master volume control either to the output of the preamp or to the input of the power amp. Whatever suits you best. Personally I would go with the latter configuration.

The output cap in LM386 -based amps (and practically in other power amps as well) has high capacitance because it decouples the amp to very low impedance. Calculate the corner frequency of the formed RC filter and you see why the value must be high. High value means that one must use an electrolytic cap. If the chip is decoupled to higher load impedance the value can be reduced a lot while still maintaining the same corner frequency. I'm quite sure the series RC "Zobel" can also be removed when the chip is driving resistive loads.

Little Gem indeed has a volume control after the chip. This is likely the arrangement one should use when coupling that kind of circuit to an external power amp. Anyway, whatever you do, I'm sure you will have to attenuate the chip's output considerably to prevent overdriving of the following stage (the power amp).

[armstrom]

Teemuk and Darwin,

Thanks for you help. Here's a partial schematic of what I plan to do (note I've only shown the signal path from input to the power amp (including the input of the power amp board) with my comments in place. This is not a complete schematic, I know I forgot the cap from pin 7 of the 386 to ground

So, what do you think? Will the 10K trimmer be sufficient to allow me to attenuate the output of the LM386 to keep from clipping the power amp with the 100K volume control at maximum? I could always add a series resistor before the 100K volume pot (maybe a 50K?) and that should increase the attenuation across the board (even when the pot is at maximum volume)

Any feedback is much appreciated as I need to get this built ASAP but don't want to waste time chasing my tail on a bad idea
-Matt

P.S. The schematic image is pretty wide, so you will probably have to scroll to see it all.

[darwindeathcat]

Looking at your schematic and also taking Teemu's wise advice into acount I can tell you the following: The "Zobel Network" is the series cap and resistor to ground at the output of the lm386. This is used to prevent high frequnecy oscillation when driving a speaker load, so you don't need it when coupling to another amplificatin stage (and Teemu confirms this in his post). He also explains that you don't need such high decoupling capacitance (the 220uf one) when coupling to high impedance loads. Your poweramp's input impedance is set by the resistors at it's input. currently you have 4k7 + 10k to ground when the trimmer is set all the way up. You'll have to calculate the output impedance of the lm386, which if i remember correctly put out about 1/2 watt of power. You'll likely need to figure out what the voltage of the output signal is (find the circuit's current draw and use Ohm's law, or just use your multimeter). Your poweramp probably want's to see an input signal of no more than 1v, so you'll need to calculate the resistance you will need to drop the signal to that voltage. Then your 100k pot in the position you have it will work fine as a VC, and you will not be overdriving the poweramp at it's most open position.
As for the new value of the decoupling capacitor, I'm not sure how to figure those calculations, so maybe Teemu or others can help you there....

Also, thanks Teemu for your advice on the FX loop idea. Of course it is a good one, and I have seen sever schematics (including in your e-book) for doing it. My major problem is that I am working with a "vintage" chassis (an old tube radio), so I was trying to limit myself to the number of holes that were already drilled in it in order to preserve the "look of the thing". So I had come up with that idea as an "unorthodox" way of doing the job while sticking to only two jacks... However, after hearing your advice, now I am considering to *gasp* drill another one myself... Hope it don't mess up the looks of the thing!

[armstrom]

Thanks for the feedback. I will probably remove the filter circuit on the output.

Just a comment about the LM386 outputting 1/2W... If my understanding is correct, amplifiers don't "generate" or output wattage. The LM386 is simply a voltage amplifier. The wattage rating is simply indicative of the maximum current that the amplifier can pass safely. If at a particular input voltage the 386 chip is outputting 1V through a 4Ohm load then the wattage is 0.25W..

One thing I want to understand though... For a fixed input voltage (say .1 V) will the output voltage be the same (2V assuming Av = 20) regardless of the impedance of the load being driven? I ask because on the LM386 datasheet there is a graph showing maximum peak-to-peak voltage vs. supply voltage and there are four curves, one for 4Ohm, 8Ohm, 16Ohm and infinity... Does this just indicate the maximum possible output voltage at those load levels? In other words, the point at which the output will clip.

So other confusions aside... The question still stands... does the schematic I posted seem reasonable? Will the 10K trimmer pot (after the volume pot) be sufficient to attenuate the signal enough so as not to overdrive the power amp?

Thanks again to everyone who has responed. I really appreciate it.
-Matt