Hi!
I'm attempting to construct the tried-and-true DIGI-125 circuit and I'm experiencing some weird behavior that I hope someone can help me with. The concrete problem I have is that I'm seeing no signal at the Q3 (VAS) base using my oscilloscope. I can see the input signal present at the Q1/Q2 collectors, but there's nothing but a steady -v9.25V DC on the emitter of Q1 when being run from ±10V rails. I have run the circuits on higher voltage rails as well, but this makes no difference. I strongly suspect that I have wired something wrong or have been sloppy in my soldering, but I've now gone through the circuit with a magnifying glass two nights in a row and double checked the transistor pinouts a number of times and still haven't found what is causing these problems. What do you suggest I do to pinpoint the issue?
Points that might be relevant:
I'm attempting to construct the tried-and-true DIGI-125 circuit and I'm experiencing some weird behavior that I hope someone can help me with. The concrete problem I have is that I'm seeing no signal at the Q3 (VAS) base using my oscilloscope. I can see the input signal present at the Q1/Q2 collectors, but there's nothing but a steady -v9.25V DC on the emitter of Q1 when being run from ±10V rails. I have run the circuits on higher voltage rails as well, but this makes no difference. I strongly suspect that I have wired something wrong or have been sloppy in my soldering, but I've now gone through the circuit with a magnifying glass two nights in a row and double checked the transistor pinouts a number of times and still haven't found what is causing these problems. What do you suggest I do to pinpoint the issue?
Points that might be relevant:
- I'm running BC556's for the long tail pair.
- I'm running darlingtons (TIP122/127) for the output pair instead of the discrete driver + output device configuration from the schematic.
- I'm using 1N4148 diodes for the output biasing.
- I have tried both with and without C3, makes no difference.
- Bias voltage on the NPN output device (no signal applied) is 0.8V and -1.1V on the PNP output device.
- DC offset at the output is +20mV.
- I observe distorted signal on the bases of the output devices and there's also a distorted output from the amplifier.
- I have checked Q3 (suspecting it was bad), but checking base-to-collector and base-to-emitter using my multimeters diode mode reads ~850mV for both, so no shorts.
- All transistors are genuine (sourced from Farnell).
Hi,
You shouldn't expect superlative performance from this circuit. That said, the voltage you report at Q1 emitter sure doesn't sound correct.
If you're seeing -9.25V at Q1 emitter, confirm same voltage at emitter Q2, and the corresponding lead of R2. Measure the R2 voltage directly on the resistor lead, near the resistor body. In like manner, +Vs should be present on the other lead of R2. If things don't start to make sense, would you report the voltages seen at all three terminals of both Q1 and Q2? Try to measure directly at the device terminals if possible to discover open connections.
BTW, good practice would have C3 present and there also should be 100nF caps at +Vs to ground, and -Vs to ground.
Good luck!
You shouldn't expect superlative performance from this circuit. That said, the voltage you report at Q1 emitter sure doesn't sound correct.
If you're seeing -9.25V at Q1 emitter, confirm same voltage at emitter Q2, and the corresponding lead of R2. Measure the R2 voltage directly on the resistor lead, near the resistor body. In like manner, +Vs should be present on the other lead of R2. If things don't start to make sense, would you report the voltages seen at all three terminals of both Q1 and Q2? Try to measure directly at the device terminals if possible to discover open connections.
BTW, good practice would have C3 present and there also should be 100nF caps at +Vs to ground, and -Vs to ground.
Good luck!
I think you are mixing up the collectors and emitters on your input transistors. The emitters are the terminals with the arrows that are connected to R3, and the collectors are connected to R2 and V-.
The -9.25V you are seeing on the collector of Q1 (connected to R2 and the base of Q3) sounds about right for +/-10V rails. The AC signal on the base of Q3 is going to be pretty small, so it is possible it is there and you just aren't seeing it.
I think the main problem you are going to run into is that the component values in the circuit were selected for a +/-37V supply, and +/-10V is too far from that to expect the circuit to function correctly. I would try running it at something closer to the design voltage and report back with some new voltage readings.
The -9.25V you are seeing on the collector of Q1 (connected to R2 and the base of Q3) sounds about right for +/-10V rails. The AC signal on the base of Q3 is going to be pretty small, so it is possible it is there and you just aren't seeing it.
I think the main problem you are going to run into is that the component values in the circuit were selected for a +/-37V supply, and +/-10V is too far from that to expect the circuit to function correctly. I would try running it at something closer to the design voltage and report back with some new voltage readings.
http://www.iki.fi/no/DIGI-125_low-res.pdfThe original magazine article is linked above and whilst its an old and poor scan, there should be some understanding there, that the specified supply is required to make the amplifier work properly for intended 50-75 or 100-125W applications and that means keeping to the original power supply voltage and circuit voltages too. If you only need a small output capability such as you might get from a +/-10V supply, a very different type and scale of design would surely be a better plan.
Last edited:
Thanks for your insightful comments @BSST, @Rory Christ and @Ian Finch! I had indeed mixed up the terminology as @Rory Christ points out and I was seeing -9.25V on the collector of Q1, not the emitter, sorry. 🤦♂️
The reason I was running the circuit on ±10V rails was to reduce the chances of blowing up any components in case I had wired something up wrong. I've now worked up my courage and adjusted the rails to ±30V (as far as my adjustable power supply will go) and I've taken some more complete DC voltage measurements with the circuit at idle.
Some observations:
The reason I was running the circuit on ±10V rails was to reduce the chances of blowing up any components in case I had wired something up wrong. I've now worked up my courage and adjusted the rails to ±30V (as far as my adjustable power supply will go) and I've taken some more complete DC voltage measurements with the circuit at idle.
Some observations:
- R6 and Q3 both get very hot.
- In-circuit with no voltage applied I measure R6 to 1.5k.
- The circuit draws about 0.04A from both rails at idle, according to my power supply.
- I removed R6 from the circuit and it measured dead on 2.2k out-of-circuit. I replaced it with a 2W metal film resistor just to be sure. No changes in voltages were observed after the change.
Last edited:
With the voltage you report, R6 would be be dissipating about 0.8W, so yes, it would be getting hot.
But something is very wrong. Roughly speaking, Q7 base should be -0.6V, and Q4 base -1.2V. Would you report what you observe? Any load applied to the output?
P.S. Make sure the correct devices are installed at Q4 though Q7. A guess on my part is that Q4 and Q7 may not be conducting any current and that Q5 and Q6 may be in reverse breakdown and providing the conduction from the output to the voltage you noted on Q5 base.
But something is very wrong. Roughly speaking, Q7 base should be -0.6V, and Q4 base -1.2V. Would you report what you observe? Any load applied to the output?
P.S. Make sure the correct devices are installed at Q4 though Q7. A guess on my part is that Q4 and Q7 may not be conducting any current and that Q5 and Q6 may be in reverse breakdown and providing the conduction from the output to the voltage you noted on Q5 base.
Last edited:
I had an 8 Ohm speaker connected to the output - removing that and measuring at the amplifier output makes everything behave as expected! 😳
With no speaker connected I measure 0.4V at the NPN darlington output transformers base and -1.8V on the PNP side. My speaker measures 8.5 Ohm using my multimeter, so I don't think there's anything wrong with that. Just to confirm I connected a 4 Ohm dummy load and with that in place the bases are back at -13V and -15V again... 🤔
With no speaker connected I measure 0.4V at the NPN darlington output transformers base and -1.8V on the PNP side. My speaker measures 8.5 Ohm using my multimeter, so I don't think there's anything wrong with that. Just to confirm I connected a 4 Ohm dummy load and with that in place the bases are back at -13V and -15V again... 🤔
Something wrong with Q4 and/or Q7. Try a moderate load on the output, perhaps 1k. I think you'll see marked deterioration as you introduce heavier loading.
I keep forgetting Q4/Q7 is a Darlington in your implementation. You shouldn't see base drop more than about -1.4V re the emitter terminal. An open failure?
I keep forgetting Q4/Q7 is a Darlington in your implementation. You shouldn't see base drop more than about -1.4V re the emitter terminal. An open failure?
Last edited:
Those darlingtons have an internal resistor between the base and the emitter, so if one or both of them is installed backwards (base and emitter flipped), has failed open, or is missing the collector connection, the VAS would still be driving the output through the internal resistor and it looks like it is working until you connect the load. For example, here the upper transistor is missing the collector connection/failed open and the lower transistor is reversed. I've been trying to figure out a scenario that would result in the voltages you are showing, but so far, I haven't had any luck.
I have drawn up a simulation for the circuit using discrete transistors. I substituted MJL2194 and MJL21193 for 2N3055 and 2N2955 since .models for the latter are not in the LTSpice onboard library. The THD at 1kHz is OK at high power mostly around 0.1% at this and other levels.
There is a snag however - at 20kHz there is evidence of slew rate limiting.
The current passing through R3 at 4k7 of around 150uA is not enough to drive Q4 with the amplified collector base capacitance together with that of the parallel150pF compensation capacitor loading the base of Q4. At around 2V peak out THD is about 1.3% and this gets worse with higher peak output voltages. I have moved the compensation capacitor to a position where there is less liklihood of slew rate limiting.
Slew rate limiting was a hot subject for debate in journals like Wireless World in the 1970's. The DIGI-125 appeared in ETI magazine in 1989 so what was going then.
Anyway I think it likely that Q3 has failed and that it is a mistake to get around using discrete driver transistors by using Darlington packages.
The difficulty with these in containing two forward biased base emitter junctions, which gives rise to a problem of heating up and compounding the passing more current.
It would be better to have the output devices on a common heat sink and have the string of diodes in contact with the heat sink. That would be awward with 1N4148 and it would be better to use something like a 1N400X style devices sitting in holes drilled through the heat sink.
I have attached a revised version of the circuit with BD139/140 which should stay cooler than the original devices and outputs MJL21194/93 claimed to give low THD even with unmatched pairs.
There is a snag however - at 20kHz there is evidence of slew rate limiting.
The current passing through R3 at 4k7 of around 150uA is not enough to drive Q4 with the amplified collector base capacitance together with that of the parallel150pF compensation capacitor loading the base of Q4. At around 2V peak out THD is about 1.3% and this gets worse with higher peak output voltages. I have moved the compensation capacitor to a position where there is less liklihood of slew rate limiting.
Slew rate limiting was a hot subject for debate in journals like Wireless World in the 1970's. The DIGI-125 appeared in ETI magazine in 1989 so what was going then.
Anyway I think it likely that Q3 has failed and that it is a mistake to get around using discrete driver transistors by using Darlington packages.
The difficulty with these in containing two forward biased base emitter junctions, which gives rise to a problem of heating up and compounding the passing more current.
It would be better to have the output devices on a common heat sink and have the string of diodes in contact with the heat sink. That would be awward with 1N4148 and it would be better to use something like a 1N400X style devices sitting in holes drilled through the heat sink.
I have attached a revised version of the circuit with BD139/140 which should stay cooler than the original devices and outputs MJL21194/93 claimed to give low THD even with unmatched pairs.
I'll reiterate--- your schematic in post 5 shows more than -15V base to emitter on the PNP Darlington. This is not possible with a functional device. It's either damaged, miswired, or not as claimed in package label.
Exactly. Single base emitter voltages will never be more than about .7 volts forward biased (what you should have).
The signal at the base of the VAS will be lower than the amp out by the gain of that stage. If the gain is 60db a 1v out only needs 1mv. Looks like this amp VAS dosnt have much gain with that 2.2K load so you will see a signal at the vas base when you get it working
The signal at the base of the VAS will be lower than the amp out by the gain of that stage. If the gain is 60db a 1v out only needs 1mv. Looks like this amp VAS dosnt have much gain with that 2.2K load so you will see a signal at the vas base when you get it working
That was exactly it - thank you so much! It was a stupid error on my side, much like I suspected all along. I had the base and emitter pins on the PNP output device (TIP127) mixed up. After flipping the base and emitter to their correct pins, everything works as expected. I will attach the output devices to an appropriate heat sink and report back with some readings.
Thanks again, all who helped! 🤗
Some observations after using this as a guitar power amp for a while:
- Plenty loud for home use running on ±20V rails into a Weber Ceramic Signature 12" speaker (8 Ohm).
- Clean enough for guitar use (see 1W screenshot below).
- I get some hiss when running the amp "open" on my bench (see image below), but nothing critical. Should clean up nicely when built into a shielded chassis.
- A Boss GE-7 graphic equalizer pedal works well as a preamp.