I actually have one of the almost-famous Kay T-1 tremolos. At Nightingale's request, I measured some voltages and did some general poking around to help out those of you that are cloning this thing.
Voltages - measured with 8.22V battery voltage and oscillator disabled (by breaking the connection between the 33 Ohm resistor and the caps)
Q1 E=0V B=0.697V C=7.09V
Q2 E=0.028V B=0.541V C=0.029V
Q3 E=0V B=0.548V C=0.029V
Q4 E=0V B=0.697V C=1.20V
SOME VERY IMPORTANT NOTES:
1. The transistors are 2SC828R, which have pinouts of E-C-B (left-to-right, looking at label). The GGG layout is for transistors with a E-B-C layout (such as 2N5088, etc.)! 2SC828 (or NTE199) will not work in the GGG board unless you twist the leads around!
2. I don't have an original datasheet for the 2SC828, but the NTE199 (it's substitute) is a high-beta (>400) part. So the 2N5088 may be a good sub that works directly in the GGG layout. HOWEVER, the actual betas of the transistors in my unit seem to be in the 100-200 range, based on the measured voltages. It might be worth checking out some lower-beta transistors as well and see what works better.
3. There is an error in the GGG schematic. The control pot is not a 1K unit - it's 5K (labelled and measured). I am the original owner of this unit, and it has not been changed. Using a 1K would give only the faster range of speeds. However, the oscillation cuts out if the pot is set above about 4.2K. So, you could put a 33K resistor in parallel with the pot to limit the resistance. In the original pedal, the pot rotation is constrained by the mechanical design of the pedal. By the way, the schematic label on the unit identifies the pot as "1Kx5" , which is a strange notation and the probable source of the error.
SOME MOD SUGGESTIONS: The first stage is not well-designed, as it is not stable with beta and temperature. It needs a bypassed emitter resistor to stabilize the bias point. Try a 390 or 470 Ohm in parallel with a 22uF cap, and then adjust the value of the 470K resistor to set the collector voltage to about 4.5V.
The bass response is weak - increasing the 0.1uF input cap to 0.47uF and the 0.01uF to 0.1uF might help, but I haven't tried it.
(http://www.snowdog.org/pedalpix/exterior.jpg)
(http://www.snowdog.org/pedalpix/interior.jpg)
(http://www.snowdog.org/pedalpix/pcb2.jpg)
thanks alot brett~
turns out that one of the 4.7uf caps i was using was shot... :? this was the first time i have ever come across a bad cap... so it threw me for a loop... i wouldn't have figured it out if you hadn't posted your voltages tho...
i'm really glad i didn't give up! once i got the cicuit going, i had 2n5089 tranny's in the sockets... the sound was really distorted, and much louder that unity... so i tryed some 2n2222's. the output was too low, and the trem effect was slight... i tried loads of different tranny swaps, but ended up with a higher hfe 2n2222 for Q1, and Q4...(around 230) and then dropped the high gain 2n5089 in the twin oscilators(Q2,Q3)...
i ended up with a very clean trem sound at perfect unity... but super
choppy... the trem decays very nicely with the note... now i understand why the originals are worth a few lute-cakes...
i am happy with this trem, and dont even plan to do the simple depth mod just leave it a one knobber... but an expression pedal would be nice to control the speed... i might have to browse ebay for one...
brett, thanks again for taking the time to get those voltages...
also~
cool to see the original too!
Sorry to be so late getting in on this thread. Looking at this schematic: http://www.generalguitargadgets.com/v2/diagrams/kaytrsc_original.gif
Brett, thanks for all the information and pictures. I don't understand your description of improving the first transistor stage, are you saying to add a small resistor off the emitter to ground on Q1 then adjust the bias?
I've had a newer version of this project in the wings for a while, haven't gotten around to finishing it and posting it yet. I think the basic circuit really is a dog...very hard to get a good setup with no distortion in my building experience.
Judging from the voltage in/out of those electros, I'd say you're right, the schematic has them backwards. I don't hear any difference in either orientation, so I guess it may not matter much, but I will correct the schematic on my site. I'll also correct the pot value to 5k, this make more sense, and sounds better.
I ended up with 2N5088 in Q2 & Q3 and 2N2222 metal cans, with very low hFE in Q1 and Q4. This gives the most intense tremolo with the least amount of distortion. Even with this, I get some distortion when I bang the low E string real hard. But as the hFE goes up on Q4 the distortion increases. I guess there ought to be a circuit tweak that might fix that. Any suggestions anyone?
Putting in 2N3904 transistors in Q2 & Q3 actually makes it sound more like a standard amp tremolo and not a volume chopper, but I like the volume chopper sound.
Thanks,
JD
>>>I don't understand your description of improving the first transistor stage, are you saying to add a small resistor off the emitter to ground on Q1 then adjust the bias?
Yes. Put a 330 Ohm between Q1's emitter and ground. Put a 22uF in parallel with the resistor. Then adjust the bias to set the collector to about 4.5V. This change will make Q1's bias MUCH more stable and provide a consistent amount of headroom on the first stage
Good luck with it - this is a really cool tremolo!
i love mine~
it's on my board... i also ended up with a really low tin can 2n2222 for Q1.., and 2n5089's in the Twin T oscilator spots...my result is a slightly loder than unity tremelo sounding effect...(not too choppy)... but it cuts through the mix really well... it also gets along with my other effects pretty well... i didn't bother doing the "depth mod" just a one knobber...
definately a cool effect, but it does need a bit of tweaking... i guess i wouldn't be into this hobby if i had a problem with tweaking things... lol!
Hi,
I'm a beginner also building something based on this tremolo and it seems we all encouter the same difficulties with the amplification stage that create distortions. I tried to stabilize the first stage by a resistor and a capacitor in parallel on the emitter and then adapting resistor values... It already sounds better, but I still have distortion on very loud attacks. My poor electronic knowledge make it difficult to improveit further.
1) One of my "plans" was to simply recreate totally this first stage with another transistor or op-amp and recalculate everything to have a better design. But to do this I need to know the gain of the original circuit... As the formulas are generally -RL/RE, I'm stuck because there is simply no resistance on the emitter of the original circuit... I imagine the gain is very high, but don't know how to estimate it. When hacking around, I was able to get something not too bad with 800R (and capacitor) on the E, and 4K3 on the collector... If what I read was correct, I can estimate the gain to this to 50. Is it correct ?
2) If correct, why is the gain of this stage so high ? I don't hear anything "loud" at the end of the circuit. Does this means the next stage (the two transistors) always drop a very large part of the signal to the ground, whenever we are in the oscillation process ?
3) If my values are still correct, is it simply *possible* to get such a high gain (around 50) with a 9V battery and an input signal from a guitar ? This looks suspect to me.
4) One last comment, from what I measured on my circuit (and what was noted on an original tremolo at the beginning of this topic), it seems the circuit sounds the best (with less distortion) when the voltage between collector/emitter is between 7 and 8 volts... I don't really understand, as everyone seems to agree that a correctly biased circuit should show 1/2 Vcc on its collector.
Many thanks in advance, Nicolas
Hi,
Still haven't really solved that problem, but I continue experimenting with this circuit, and I will write my notes there so various infos about this circuit are available here :
1) Discovery of the day is that my tremolo works and sounds MUCH better when I put the "Don Tillman preamp" (basically a very simple FET buffer with 3db gain) in front. I suppose the input impedance of the kay tremolo is bad and this preamp solve this... Difficult to describe in details, but bass are preserved (they were not), the sound is more present, more defined, and I have the feeling the effect mix better with the guitar sound...
2) I was able to make distortion almost inexistent by lessening the gain of the first stage amp (see my previous questions). I simply put a trimmer in series with the capacitor that is on the emittor. Of course as I make the gain lower, I really feel a large attenuation in my level when I put the effect 'on'... I suppose I should then make another amplification system on the output of the effect ? To be investigated...
3) I have an issue since the beginning with this project, I can hear the 'oscillations' when I don't play and there is silence... I suppose the are some kind of interferences, but don't know how to look first... I use good shilded cables to connect my PCB to jacks and pots. If I choose to boost the output gain by adding an amplification stage (explained at end of point 2) ), I'm afraid I will amplify these undesirable oscillations too. They're not "huge", but we can clearly hear them. Maybe they could disappear once my circuit is in a mzetallic stompbox ?
Thanks to anyone interested in this fun circuit !
Quote from: niconoe on February 05, 2011, 06:20:59 PM
2) I was able to make distortion almost inexistent by lessening the gain of the first stage amp (see my previous questions). I simply put a trimmer in series with the capacitor that is on the emittor. Of course as I make the gain lower, I really feel a large attenuation in my level when I put the effect 'on'... I suppose I should then make another amplification system on the output of the effect ? To be investigated...
Yes, the distortion is the biggest problem. In the post above from jsleep, it seems he solved the problem:
"I ended up with 2N5088 in Q2 & Q3 and 2N2222 metal cans, with very low hFE in Q1 and Q4. This gives the most intense tremolo with the least amount of distortion."
So, there is probably no need for a trimmer or another amp stage...just choose transistors carefully.
You're right, that'll be my bext attempt, and I will post the results here !
I tried by using a 2N2222A (with hFE 200, it's the lowest I have) for Q1 and played a little with the resistor values, but I didn't get a real difference with the 2SC828.... Either it saturates on loud attacks, either the sound is quite attenuated (when I add more resistance on the emitter).
Anyway, after another bunch of tests I ended with something maybe not perfectly designed but that I like a lot; I'll let it as it is for the moment.
Here are the differences with the original schematics found on the web.
1) I implemented the depth mod, by replacing the 68K resistor by a 100K potentiometer.
2) I added a 27K resistor in parallel with the speed pot to limit its value (as explained in the first message of this thread, but 27K was a better value than 33K in my case).
3) I kept the original 2SC828 transistor everywhere, but I tweaked the first stage... I put a 800R resistor on the emitter, with a 22ยต cap and a 10K resistor in parallel. That stabilize the amp, and lower the gain a bit. Note that when playing with the biasing of this transistor, I replaced the 470K resistor between the base and +9V by a trimmer. I notice I have the best result when I set this trimmer to get between 7 and 8 volts on the collector. My poor electronic knowledges think it's strange, I'd bet on 4.5V there. But I tried it, and I have more distortion with 4.5V than 8V there...
4) As the level was then lower when the effect is engaged (because of the gain reduction), I put TWO "don tillman's FET preamp" on the circuit. The first is on the input, and it gives a more defined, present sound, with a better bass reproduction... I suppose it's the inpedance adaptation of this circuit that makes this effect, but it really enhance the general working of the tremolo... It also gives a 3 DB boost on the input, but it's still ok, thanks to the added stuff on the emitter. Then the second circuit is on the output, to bufferize it for further effects and gives another 3db boost. With these two boosts/buffer, I don't feel any attenuation when I enable the effect.
5) Build a simple true bypass with a triple inversor (inpout-output-LED) around this circuit.
All that together allows me to keep the "Kay choppy sound" (but more versatile due to the depth mod), but without or with very few distortion and a less "lo-fi", thanks to the two buffers... At least I really like the result, hope this notes will help someone one day...
Thanks,
Nicolas