The voltage on the input of 7915 regulator is much low. It also seems to fluctuate on the clock signal. I put 24.3V on the power input (it remains stable). After R57 there remains a voltage that fluctuates between -17.58V and -17.84V (with delay knob fully open).
When I turn the delay knob all the way down, -15.13V and -15.35V remain. In short, the voltage after R57 responds to the clock.
When I remove IC9, I am left with a stable voltage of -18.42V to R57. This one is stable.
However, this voltage is still too low if I may believe the manual. This indicates that the voltage on the input of the 7915 must be -25.57V.
So, what's wrong here. Does anyone have an idea?
(https://i.ibb.co/86Sx2KV/power-clock-section.png)
The problem seems to be with the power input. Take out all the IC's except the power regulator, and verify the following against the build document:
1. Power is wired correctly, this is a positive ground circuit which can be confusing.
2. You're using the recommended power supply.
3. When you apply power, it's the correct polarity and voltage entering the PCB + and - holes.
4. D1 is the right part number, and installed with the proper polarity.
5. R57 is the correct value.
6. Then check the voltage at the regulator input and output.
The dropout voltage of the 15V regulator is about 17V input, so if that's your input voltage, it's on the edge of dropping out, that's probably why the power is responding to the draw of the circuit.
Yup, mauman is spot on. You don't need more than -18v or slightly less for the LM7915. The main concern - is the regulator voltage actually fluctuating, or just the supply voltage to it? I do wonder what might be causing such a fluctuation...perhaps current draw is itself causing an issue.
Thanks for the quick responses.
The power supply is a well-known (by Dutch standards) Delta Elektronika E030-1. This power supply is known for its good performance. I'm assuming that's not the problem here.
The voltage on one side of resistor R57 remains constant (-24.3V). There is an error on the side connected to the zener diode and the LM7915.
What is the reason for this zener? At a voltage of -24V, this zener will still do nothing, since it has a working voltage of 30V. Doesn't this protect the LM7915 only if the offered voltage is higher than 30V?
Could I remove it for the test?
I will take pictures of the test setup tonight and measure the whole thing without ICs.
Another question. The wires from the chorus/vibrato switch run along the power supply wires. Could this also be of some influence?
Okay, here we go. Here are some measurements. To start with all ICs removed.
Then the voltages are stable and appear to be correct.
Then I put IC9 back. Then the voltage fluctuations appear on the input pin of the 7915. The voltage after R57 also drops a bit. However, the voltage is still so high that the voltage on the output of the regulator remains stable at -15V. Then I keep putting an IC back and do another measurement. The voltage at the input of the regulator continues to fluctuate and drops downwards. Until all ICs are inserted. From that moment on, the voltage has dropped so far that the output of the 7915 also fluctuates.
Finally, a measurement was made with all ICs except IC9. Then the fluctuating is gone. Voltage has still dropped too far. Then another test except with the IC8 and IC9. Then you can see that the voltage is rising again.
Basically, the voltage drops as the number of ICs increases. And IC9 ensures, among other things, a varying voltage on the input of the regulator.
(https://i.ibb.co/9WdthQs/Measured-without-IC.jpg)
(https://i.ibb.co/NKyWnK2/Measured-with-only-IC9-inserted.jpg)
(https://i.ibb.co/vzcwnqD/Measured-with-IC8-IC9-inserted.jpg)
(https://i.ibb.co/XZyMzjk/Measured-with-IC1-IC8-IC9-inserted.jpg)
(https://i.ibb.co/wB0N7Xk/Measured-with-IC1-IC2-IC8-IC9-inserted.jpg)
(https://i.ibb.co/XSM2xZp/Measured-with-IC1-IC2-IC3-IC8-IC9-inserted.jpg)
(https://i.ibb.co/4m6pxmp/Measured-with-IC1-IC2-IC3-IC4-IC8-IC9-inserted.jpg)
(https://i.ibb.co/WcPhRmz/Measured-with-IC1-IC2-IC3-IC4-IC5-IC8-IC9-inserted.jpg)
(https://i.ibb.co/D8Y9Vs3/Measured-with-IC1-IC2-IC3-IC4-IC5-IC6-IC8-IC9-inserted.jpg)
(https://i.ibb.co/X7jhJ3z/Measured-with-all-IC-inserted.jpg)
(https://i.ibb.co/4FjHz60/Measured-with-all-IC-inserted-except-IC9.jpg)
(https://i.ibb.co/BqzqWgk/Measured-with-all-IC-inserted-except-IC8-IC9.jpg)
Quote from: mauman on November 23, 2021, 02:29:22 AM
The problem seems to be with the power input. Take out all the IC's except the power regulator, and verify the following against the build document:
1. Power is wired correctly, this is a positive ground circuit which can be confusing.
2. You're using the recommended power supply.
3. When you apply power, it's the correct polarity and voltage entering the PCB + and - holes.
4. D1 is the right part number, and installed with the proper polarity.
5. R57 is the correct value.
6. Then check the voltage at the regulator input and output.
The dropout voltage of the 15V regulator is about 17V input, so if that's your input voltage, it's on the edge of dropping out, that's probably why the power is responding to the draw of the circuit.
1. Power is wired correctly, this is a positive ground circuit which can be confusing.
2. You're using the recommended power supply.
3. When you apply power, it's the correct polarity and voltage entering the PCB + and - holes.Yes, see attached image
(https://i.ibb.co/dPSttLF/IMG-20211123-214616.jpg)
4. D1 is the right part number, and installed with the proper polarity.Yes, a 30V zener diode
5. R57 is the correct value.Yes, 240R
(https://i.ibb.co/Btn1NfZ/IMG-20211123-214748.jpg)
6. Then check the voltage at the regulator input and output. They were discussed in my previous post
Quote from: Ricotjuh on November 23, 2021, 09:55:22 AM
What is the reason for this zener? At a voltage of -24V, this zener will still do nothing, since it has a working voltage of 30V. Doesn't this protect the LM7915 only if the offered voltage is higher than 30V?
Could I remove it for the test?
Yes, it's just there to protect the LM7915 from too-high supply voltage.
Quote from: Ricotjuh on November 23, 2021, 09:55:22 AM
The power supply is a well-known (by Dutch standards) Delta Elektronika E030-1. This power supply is known for its good performance. I'm assuming that's not the problem here.
This power supply is capable of current limiting. Is it possible the current limiter is set below the operating current of the pedal? Try setting the current limit to the maximum (1A) by turning the Current knob fully clockwise, and the pedal will only draw as much current as it needs. Set the voltage to -18V or -20V as Bean said.
I also tested at -20V input voltage, but the problem still persists. There is a voltage drop across R57 that puts the input voltage of the LM7915 on the edge to output a -15V. I think Brian means the regulator input with the -18V. And not the adapter voltage. Otherwise R57 will have to be adjusted I expect. Well, the problem still persists. For the following tests I have set the supply voltage back to 24V
I have the current limiting fully open. So the full 1A can flow there. But only 40mA is consumed.
I measured with my oscilloscope. Here you can clearly see that after R57 a ripple dances on top of the DC voltage.
(https://i.ibb.co/m5TStc9/PRINT-10.png)
This is measured on the incoming power supply.
(https://i.ibb.co/qnhbNKG/PRINT-08.png)
This is measured at the input of the LM7915. Here you can clearly see a pattern showing the fluctuating. Given the period time, this seems to flow from modulation/clock signal?!
(https://i.ibb.co/4Ktd4Dt/PRINT-09.png)
And finally a screenshot of the measurement at the incoming power supply and the input of LM7915. Then the difference is clearly visible.
Seems like the LFO is inducing some noise with the pulses.
Why use a 4558 as LFO? They are the worst ICs for LFOs. Test it with a TL062.
I never understood why they used a 24V supply for a pedal that basically runs at 15V. Were regulators that bad in the 70s?
Here is question if effect works wrong way? If not, here is only normal situation related to power consumption of effect circuits, that reflects in current consumption and it logically reflects in power "swings", that you can see at measure points. If any current gets higher, at resistor in pathway you will see voltage changes. Nothing surprising and nothing wrong.
Try jumping around the 240R and reduce the supply voltage to 18V or so. If E = I * R, won't 40 mA through a 240 ohm resistor drop 9.6V? To test, try lowering the supply voltage to 18V and jump around the 240R. If that's satisfactory, decide what your supply voltage will be, and pick R57 value to suit. A 100R should only drop 4V and dissipate 160 mW, so 24V in = 20V at the regulator.
Output at stabil -15V is only important. And how the power source is strong and filtered of course.
BTW, the noise pulses are not relevant to the low-input-voltage issue. Regulators are not good filters, and the LM7915 is just passing along the noise that's coming back up the -15V bus from the ICs.
Could it be that in the original, the input voltage is higher so the problem doesn't occur? In combination with the 240 ohm resistor. Say the input of the 7915 must have -18V. At 40mA, that means the incoming power supply should be -27.6V. Now the Memory Man comes standard with a 24V adapter. Now it is often the case that the voltage with an adapter without load is higher than the specified value. And as the load is, the voltage will also drop slightly. But is it plausible that this voltage is higher than the supply voltage I set with my stabilized power supply? Perhaps with a "real" 24V, the resistance value of R57 will have to be lowered.
I also ordered an original adapter which I will receive next week. I'll see what the power supply does then.
Just received the adapter. Without load, it therefore supplies more than 30V.
Once connected, it drops to -27.90V.
Then I am left with a voltage at the input of the 7915 that fluctuates around -20.80V.
A stable voltage of -15.14V remains at the output of the 7915.
Next I want to measure and compare the voltage as shown on page 14 of the manual.
If I then measure on pin 1 of ic2, it continues to fluctuate. Just as mentioned before as it does for the regulator. But at a different voltage. These voltages are not stable.
And this is with a lot of pins.
How have others experienced that?
I believe that this power branche is ok. BTW if power feed at input is about 20V, at stabil is lost 5V, to get 15V. With current 40mA at stabil, those 5V will give 200mW of junk power lost in heat at stabil. Good cooler will make it up. Important is that this stabil will eat some power drop, to work properly, so you must put at input something more than outpot voltage. 20V is ok. Higher voltage drop at input-output at stabil - higher junk heat.
Yes, I also believe that the power supply is fine now. I also have a stable -15.14V after my 7915. So that's good.
But I've added a video, and there you can see why I still have my doubts. As an example I measure here on pin 3 of the MN3005. Now according to the manual I should measure -9.71V. But as can be seen in the video, it is not stable. This also applies to some other voltages.
https://youtu.be/f6O5aHMeq9Y
Everything looks correct. Is the effect no working properly? I believe that one has a bipolar cap, doesn't it?
Sent from my iPhone using Tapatalk
Quote from: jimilee on November 30, 2021, 10:52:21 PM
Everything looks correct. Is the effect no working properly? I believe that one has a bipolar cap, doesn't it?
Sent from my iPhone using Tapatalk
Yes, it has a bipolar capacitor in it. C35, 2.2uF and and is in the feedback loop of IC9_B.
I have not tested the pedal yet. I want everything to be set up correctly first. So you still have to set trimmers. I wanted to do this with a function generator and oscilloscope. But unfortunately I can't find any instructions for this on Madbean. I found the link to David Morrin on how to calibrate the EH 7850. However, I don't think that description is entirely clear.
So, does anyone here have a step-by-step description of how to set up the trimmers with the oscilloscope?
Edit:
Just trying to calibrate with an audio probe. Followed instructions as stated in the manual. Then connect audio probe to either pin3 or pin4 or IC6. Then I hear a continuous high-pitched beep. When I measure this with my oscilloscope, I see the following picture. Clock signals look good. There is no signal at the input. But I do see a signal around 8.6 kHz. (On pin 4 this is also the same but I see that my measuring cursor is not in the right position, so it does not measure the full period here)
Pin 3 and 4 do differ in amplitude. I should only measure my delayed signal here, right? In this case I shouldn't measure anything as I have no input connected.
Also tested with input signal. Then I indeed hear a delayed signal on pins 3 and 4. But the high frequency predominates.
Please your expertise.
Depending what you are trying to achieve you need to have a signal or not.
To bias the BBD use a sine wave, increase the gain so that the wave distorts in the BBD and adjust the bias so that the wave is roughly symmetric. This depends on the delay setting, so either choose your preferred setting or set it with the pot half-way. You want to have the same headroom on both sides, using the distorted wave helps, because it definitely uses all the headroom available. The clock noise right after the BBD is normal. I prefer to measure at the input of the opamp, so that it is away enough from the output and some of the noise is gone. Do the same for the second BBD.
The gain between BBDs should be adjusted so that the second BBD sees the same voltage as the first.
To adjust the balance at the output of the second BBD I do it without signal. You will see the clock noise and you can minimize it by adjusting the trimmer. I am such a fan of this thing, that in my build I added a trimmer right after the first BBD too. I don't think it helped though, the noise level seems to be the same.
The gain stage after the second BBD is adjusted so that you have unity gain between the two sides of the compander, if I am not mistaken.
It is basically summarized here, although here you have different BBDs. You will have to try different delay settings until you find one where most of the delay time is usable. For the noise balance, the highest delay time is suggested, since everything above 20k is not audible anyway.
(https://sites.google.com/a/davidmorrin.com/www/_/rsrc/1462497222354/home/trouble/troubleeffects/electro-harmonix-memory-man/eh-1307b/1307B%20Layout.PNG)
Just took a measurement. I still had a BBD Probe from D_H. So I did the measurements with that.
Output signal from the first BBD.
(https://i.ibb.co/Jq5pfpL/PRINT-40.png)
Measured at the same point but raised the amplitude of the input signal. Here it can be seen that the signal is starting to clip. With BIAS 1 adjusted in such a way that it is cut off equally on both sides.
(https://i.ibb.co/bBHpLvP/PRINT-41.png)
The same point here, but with an even higher overload.
(https://i.ibb.co/yPC9b8Y/PRINT-42.png)
The following points were measured after the second BBD. Same procedures here too.
Here's a clean signal on the output.
(https://i.ibb.co/dDvcGZJ/PRINT-43.png)
Slightly distorted here.
(https://i.ibb.co/M9nD3Mb/PRINT-44.png)
And here it is again distorted. But still neatly symmetrical.
(https://i.ibb.co/8YYQwrG/PRINT-45.png)
So I think the bias adjustment is done right?!
Then there are the gain settings. This seems to go well for the first stage. Here I get amplification set to 1x between pin 7 of ic6 and pin 7 of ic7 by means of trimmer Gain 1.
But then Gain 2. So when I go through the next step: "Connect your audio probe to pin1 of IC4_A. Adjust the GAIN2
trimmer until the output is the same or slightly above the volume input of pin7 on IC6." Then here's a difference. Where I measure 3.28V on pin 7 of ic 6 and 1.28V on pin 1 of ic 4.
But when I measure between the NE570 compander on pin 7 and on pin 14, I get it properly adjusted to a gain of 1x by means of the Gain 2 trimmer.
Which of the above methods is correct?
Hi,
Bias seems OK.
According to Howard Davies for the EH7850: "TRIM 5 (GAIN 2): Unity between NE570 pins 14/15 and 7.
Set the 2nd gain trim last looking at pins 14 and 15 of the NE570 and setting this trim so that the signal is equal in level to that at pin 7 of the 570. If this can't be done, go as far as the trimmer permits and re-adjust the first gain trim to achieve it.
"
I would use this value, since it will give you unity over the compander.
I guess you could also adjust for unity between IC1_A pin 1 and IC5_B pin 7, both sides of the blend knob, so that you have a perfect blend when the pot is at 12 o'clock.
Another aspect is, if you want the pedal to oscillate at max. feedback. You can adjust gain 2 so that there is oscillation, or not. Your choice.