HiwattDr103 pedal clone unstable

[Sonny]

I finally like to post a project I'm running a couple of months now. It's nearly done but there's still one problem left and
I obviousy don't have any clue.

It's an IC-based emulation of a Hiwatt DR103 as a (larger) pedal. It's based on this schematic here

BAJA Hiwatt DR103 amplifier pedal 9v 120318

Here is my version of the Hiwatt DR103 normal channel preamp and amplifier output stage - I have not built it as yet (waiting on some enclosures) but it should work ok :wink: cheers bajaman The master volume can be omitted if desired - simply replace it

www.freestompboxes.org

All in all it's working very fine and sounds are fantastic. But when I increase gain and/or the amount of input signal e.g. with humbucker or an additional boost in front I got an disturbing bias-like noise (fuzzy, starving, dying battery) noise on the top of the audio signal.

I already did a bunch of tests to locate it or get rid of it:
a) Checked connections to v bias in the feedback networks
b) Took the breadboard far away from other voltage sources + drove the circuit from a fresh 9V battery.
c) Swapped the TL072 and even the TL062 , changed the 2SK117, changed diodes.
d) took an audio probe on different Points of the circuit, starving noise seemed to apear right after IC1a (in signal line)
e) Read out voltages on both JFETS and ICs. Voltages an all pins -without increased- input seemed to be quite usual, all readings between 4.15-4.95V. Increasing the input per boost decreased voltage on the output of the first position 2SK117 and the inverting input of the TL072 temporary - then increasing fast to stabilised values as the guitar swungs out. But that's the usual effect of saturation via boost, isn't it?
f) suspicious on this deviation I tweaked different resistors in feedbacknetwork of the IC1a. Increasing the Feedback resistor (3,3K) increased gain but
no change in disturbing noise. Decreasing Biasing resistor in front of IC (1 M) increased input voltage - but this doesn't lead to vanish the starving bias noise at increased input levels.

So right now I'm kind of desperate. The pedal sounds great, but for completion it
is necessary to drive it in sweet terrain via boost or fuzz.
So I would be very very thankful for suggestions or hints and of course
good questions.

Best regards.

 

[Vicious C]

I’ve actually built a few of Baja’s versions of amps, and I have to say I’ve had the same issue, and I never was able to diagnose it. I hope you get some answers! I wish I had them!

 

[Brett]

I'm no electrical engineer, but I see a few areas in the power supply scheme that I'd start with:

1. No bulk cap for +9VDC. None.
2. There's no current limiting circuitry in the supply.
3. No form of reverse polarity protection whatsoever.
4. No form of power filtering in the supply whatsoever.
5. A 1µF bulk cap for reference voltage (v bias) is pretty conservative.
6. I wouldn't put money on it, but if I had to guess, I'd assume that this circuit is probably fairly noisy and may even oscillate at extreme settings, even when working as designed.

I start by establishing a filtered, current limited +9VDC supply. Something like this:


Treat the former +9V in Baja's schematic the same as VCC in the image I posted. From there, replace the 1µF capacitor at the junction of the v bias voltage divider with something more substantial, start with a 47µF.

Not saying this WILL fix the issues you're having, but even if they don't, they'll at least improve your power supply.

 

[Sonny]

At first thanks to you for the fast reaction.
@Vicious C : Thanks for your wishes, I let you know .

@Brett
I have to admit having left out some information of the final version of the circuit.
The basic schematic is in fact the posted one from bajaman. Tone stack is slightly advanced as shown here

...but that doesn't seem to touch my problem.

But for the finally wired version I went with this veroboard

BAJA Hiwatt DR103 Amplifier Pedal

Schematic and info HERE

dirtboxlayouts.blogspot.com

So if I'm not completely wrong, my actual version has got
3. a reverse polarity protection in the form of the 1N5817 (series diode version) and
4. basic power filtering in the form of a 100uF electrolytic cap from hot to ground, but not this complex as your suggestion

This is my first circuit this complex (went with transistor based circuits until now)
and I have to admit I let myself be seduced by the apparently sucessful buildings
in the comments beyond the circuit postings. So I'm open as much as I can and free to ask

1. What exactly is a bulk cap?

To 6. In fact the circuit is on breadboard and even soldered with open shielding quiet as could be. I had some parasitic oscillation in the beginning caused by my own fault - wires of IN and OUT ran to close to eachother. AND increasing treble and/or presence increases a bit of white noise - but still works for me.

5. Thanks to your suggestions, I'll try this partly. After doing audio and voltage probes, I still guess problem has to be somewhere round IC1a OR the power supply/voltage divider.
My thesis comes from this: in voltage test the voltages on all IN and OUT pins of the 4 TL062 Stages kept unchanged at around 4.5 even with increased input signal. Only the Voltage at non-inverting IN of IC1a went crazy. AND in Audio Probe (like this one https://diy-fever.com/misc/audio-probe/) the problem first appeared AFTER IC1a. So maybe theres a bug in power supply and TL072 is the only IC which reacts weird OR theirs something wrong around that biasingnetwork.

Is it possible to be just one resistor or cap - no false value but in lack of sufficient wattage...like a breakdown of a single part?

But I'll try the supply suggestions at first, thanks a lot.

 

[m4268588]

e) Read out voltages on both JFETS and ICs. Voltages an all pins -without increased- input seemed to be quite usual, all readings between 4.15-4.95V. Increasing the input per boost decreased voltage on the output of the first position 2SK117 and the inverting input of the TL072 temporary - then increasing fast to stabilised values as the guitar swungs out. But that's the usual effect of saturation via boost, isn't it?

Try measuring Vbias instead of the signal pins.
This shouldn't fluctuate because it is a buffered, but you may find something strange.

 

[Sonny]

Greetings to you. I got news and maybe something which may lead to breakthrough.
It's a bit odd, like life ways go. I thought today wouldn't be a day to take this project further as for
I went to a hike with my goddaughters family. But her father is an electronic physicist and on base of
my description of the circuit he directed my view of the problem towards a totally different point.
In his opinion the biased voltage (as served by the voltage divider) isn't that much important for the
undersupply as I may think.

Instead the crucial thing for working with IC's stable is the supply voltage on V- and V+ of the IC's.
Caused by uncalculated (and maybe uncalcuable) wire lenght and sometimes thin copper stripes of
the vero, the IC's aren't able to handle the input and most notably irregular strong and fast input peaks.
The system (sorry my inaccurate language, I'm still learning in that electro physical world) just aren't able
to manage the required power and collapse.
So a standard answer to this phenomen are support capacitors between V+ and V-.
As I could understand, they load up like a tank and serve needed voltage very fast in irregular power demands
- just like an increased input by a hot pickup or a boost.
In some way the 100uF filter cap I already got right after reverse polarity diode can be considered as something like
that. But...there still the connections on the vero which may be odd.
So his recommendation was to try to add support capaciators as near as possible between V+ and V- pins of the IC's.
This could be at least an electrolytic of about 10uF (you surely could calculate this) or combinations of different types
and values of caps just like the 100uF + 100nF in Bretts suggestion. By combinations like this you could even tweak
the reaction of the IC's to overdrive in different frequencies.

So at least for moment I've got a direction. I'll do the improved current limiting unit (thanks Brett) and try to increase the value
of the 1uF in the voltage divider (thanks again) - in next days I guess. And finally I will try to experimentate with this idea
of support capaciators and hope to see some change.

What do you think about this supproach? Does it sound conclusive enough to give it a try?
Thanks again for your support!

 

[Sonny]

Try measuring Vbias instead of the signal pins.
This shouldn't fluctuate because it is a buffered, but you may find something strange.

Thanks! Again my lack of knowledge: Where would I measure Vbias instead of at the pins?

 

[Sonny]

P.S. I guess I have to correct the term of the described capaciators. According to this blog here

How to choose capacitor for an IC

I'm a beginner in electronics from the software industry. With some self-taught things, I'm trying to implement some basic Arduino circuits. My confusion is about capacitors mainly. My understanding

electronics.stackexchange.com

they're named decoupling capaciators.

 

[Brett]

Decoupling capacitors are generally considered good practice. Place a 100nF capacitor as close as physically possible to each IC's V+ or Vcc pin. In this circuit, that means three 100nF caps, one per IC as close to pin 8 as possible. TL072 and TL062 share the same pinout:



The lack of decoupling may not be the problem, but it's another one of those things that will improve the existing circuit either way.

 

[Sonny]

All right. I'll try that. One more consideration is to think about the JFET in front of IC1a (2SK117).
As I remember the voltage probe showed (with increased input) an irregular drop at the output of
this JFET AND the non-inverted Input of the IC1a. As I understand considerations about voltage drops
on IC's they can be a lack of temporary supply voltage at an IC on the one hand. This could be compensated
with a decoupling cap.

On the other hand this is the consequence of another device in the circuit sucking extra voltage
to operate properly. I wonder if the JFET (which lies right next to the IC1a and is the first amplifing device confronted
with increased input) could be this device. This could be confirmed by the audio probe:
The JFET compensates increased input with extra voltage sucessful - audio signal right after the JFET is fine. But
this compensation sucks needed voltage from the next partner in the signal line (and nearest wired device to supply voltage)
- the IC1a. Am I thinking wrong?
If not, considerations about decoupling caps let me think about a decoupling cap right there, where extra voltage is needed in the
first place - at V++ pin of the JFET. Let me know, if that sounds conclusive or rather rubbish.

 

[Brett]

Please post some clear pictures of your build showing the front and back of the board along with all wiring and jacks. It may be easier to diagnose the issue if we can at least see what you're working with.

 

[Sonny]

Please post some clear pictures of your build showing the front and back of the board along with all wiring and jacks. It may be easier to diagnose the issue if we can at least see what you're working with.

Good point. I'll try to manage it this evening, even if I may be ashamed in point of my soldering work - not beautiful in any case.

And second - important- there are some adds in the final vero, I didn't mentioned so far, cause it doesn't seem to be important for my problem.
a) I doubled first gainstage for a real bright channel. Seem to be important -I agree- BUT in mentioned voltage and audioprobe in
breadboardphase of the building process I decoupled the second channel COMPLETELY from the main board and this doesn't effect the
described measurable and hearable phenomens. Even in actual test at the soldered vero described problem appear in either of the channels.
b) Added a switchable low pass filter in the front of "normal channel" to facilitate the use of beloved treble focused pedals as Zonk and TB 1.5.
Didn't use this option in actual tests on final board.

So pictures are coming soon. Thanks a lot for your support.

 

[Sonny]

So here are pictures and graphics of my actual version.
Base used to be this one https://dirtboxlayouts.blogspot.com/2020/02/baja-hiwatt-dr103-amplifier-pedal.html

"Schematic" of my version is this this one - sorry it's only a excelbasic graphic. That's the way I work during developement phase.



And here the live photos. It's already in the box since I thought it could be a shielding thing - after my first analysis phase on breadboard. So everything now is operation at open heart .

Front


Back

• Switchable low pass filter (like a "smooth switch") in front of normal channel in the upper right corner
• On the right side of the vero the additional Bright Channel, smooth switch changes value of coupling cap
• arrival of V+ and V- on the lower left corner, there still seem to be some place to expand current limiting device as suggested

 

[Sonny]

P.S. two photos of the overall. Wiring looks a bit chaotic actually. Thats mainly because I loosened some cable managements
to remove the routing of input and output wires from eachother. Your presumably judgement about long wires between board and pots are justified in all points. Thats due to my wish to finalise it in a wooden box. But fortunaly I don't hear any significant influence on basic tone and the thing is still very quiet.
In the analysis of my actual problem it may show if I'm still right to be pleased with this state.

 

[Brett]

I'll try to take a look. It may just be me, but I don't know how you could work from your graphic and come out with something that works. It hurts my eyes (pretty badly actually) just looking at it.

 

[Brett]

And second - important- there are some adds in the final vero, I didn't mentioned so far, cause it doesn't seem to be important for my problem.

You didn't mention this, and everything CAN matter when trying to help solve an issue. We know nothing more than what you share when asking for help. Right away, if that extra opamp is a dual opamp IC, you've broken a pretty cardinal rule about leaving pins hanging open. I see no connections to pins 1-3 on the IC to the far right and this can cause stability issues.

Since you've added circuitry to the original circuit, your problem could very well be there (if everything else is correct). Can you take some time and prepare a proper schematic showing all of the changes you've made? Parts values of the new circuitry, any parts values subbed from the original circuit, etc. are necessary here. I know everyone has their own way of doing things, but this skittles diagram ain't chewin' it for me.

 

[Sonny]

Hi Brett,
at first, I'm really sorry for your eyes. As you just said, it's the way, it works for me, in the end just a "more colorful" verographic.
I did a bunch of pedals this way, so yeah at least it doesn't seem to be harmful to me. But I will try to avoid showing it to other persons from
now right in . The open ends of unused IC was a real fault, the mentioned physicist also pointed me to this, thank you too for the hint.
So I will send the input to ground - is that the right way?

I'll do the schematic, but this will take some time. Thanks for patience and open mind.
Until that I think it's important to mention one thing:

I didn't change any values in the heart of the circuit. My electronical knowledge is just too limited to dare something like that, really.
I just sensibly changed some things like frequency shaping coupling caps and added a more cosmetique low pass filter (optional) and they worked without any - means notable, hearable consequences.
So I'm really glad about you folks and your help. And I really do think your attentive notice
that every change could be source of error is very worthy.

So I think it's important to point out (one more time) that the problem we're talking about happened to appear the first time (and from that on) in a circuit WITHOUT ANY MODIFICATION: Second channel was completely decoupled and the simple low pass filter in front came long after my first analysis of the problem. So I first tried to analyse it on breadboard with a really pure circuit (REALLY EXACTLY LIKE THIS VEROBOARD https://dirtboxlayouts.blogspot.com/2020/02/baja-hiwatt-dr103-amplifier-pedal.html ). And the phenomen still sounds exactly the same way like these days.

So I have to apologize for not having mentioned the modifications. I guess thats because I knew the phenomen came up long before any modification on the circuit. I may be wrong, but on base of this experience I still don't believe in a connection.
But know I see, I should mentioned the whole circumstances in the starting post. I'll remember it for future requests.
And I guess a less colorful and proper schematic will help for better analyse anyway and maybe I will be surprised in the end.
I'll get in touch.

 

[Sonny]

Hi folks, dear Brett,

this was long time no see and I apologize for that. There were some weeks of tight work phases and some illness too, unfortunately.
But still I guess, I can take you to the next step in my analysing journey of the problem, cause I managed to do the proper schematic (I hope) and some tests to exclude some more possible causes.

Once more: You'll see the modifications and they do what the should do - they are working fine.
The buzz noise appeared the first time in a pure breadboard design according to the original single channel design and every test now (in soldered version) is done in either one or the other of the channel and without the optional low pass filter
So here is what I did so far:

Test Phase 1 - voltage and bias voltage

• I swapped almost every part in the current limiting unit
• I upgraded current limiting unit as suggested by Brett - you'll see it in the schematic
• I swapped the 1uF in the bias voltage divider IC for higher values (47uF, 100uF) - doesn't seem to touch the problem but made the sound a bit dull, so I kept sockets in this place to decide in the end (ha)

• Try measuring Vbias instead of the signal pins.
  This shouldn't fluctuate because it is a buffered, but you may find something strange.

  Finally got what you meant. I already read the Vbias for I measured every pin on every IC - even on the voltage divider and therefore the vbias. Vbias kept stable on half the voltage input even during higher signal input.
• I swapped out every IC (TI072 and TI062) for a socket with decoupling cap (100nF).
• I tested combinations of 100nF decoupling cap with higher valued caps (22uf, 47uF) for I read somewhere their influence on deeper frequencies - but without results.
• I connected the pins of the unused IC (IC2b) as an independent voltage divider - as suggested by texas instruments. Improvement for sure - no influence on the problem.
  
  Test Phase 2 - Biasing aspects at and before first IC (IC1a)
  For I still assumed to have heard the noise after the first IC in audio probe I checked the parts, which could produce a biasing alike noise in the first way - in this first part of the signal line - for wrong biasing
• tested every resistor in feedback network of the IC1a/IC2a in value (per pot) - no results
• tested 1M before and 100 k resistors at the JFET in value (per pot) - no results
• tested to move the 1,5 M impedance resistor at front of signal line - in my wiring I had positioned this on directly at the first switch, wasn't sure if this could destabilize the JFET - no difference
  
  Test Phase 3 - Wrong part in first IC section
  After testing the bias option without results I had to follow the path of the first signal section till the end
• swapped the TI072CP I had used for some TI072IP from an other trader - got a bit more hiss, but no results in problem zone
• swapped the coupling cap in the first place (1uF) for a WIMA MSK2 foil for I had read about possible interference of MLCC caps with irregular signal levels, buzz even got worser a bit
• swapped the coupling cap after the JFET (1uF) for a WIMA MSK2 foil for the same reason - nothing changed
• swapped the 2SK117 for another 2SK117GR from another trader - no results
  
  Some other test, I've done: run it per battery again, run it with other guitars and cables again, run it in a completely different room

So this is my actual state of excluding issues. I have no clue again, but I'm still laid-back and don't mind to dump it down. I hope my tests may help for your fresh view on the schematic. With best regards, Sonny

P.S. I did a recording, but it's not accurate. In real (ear)life the natural gain structure of the preamp is more present - the recording seems to pick out the problem noise in particular.

 

[Brett]

Thanks for providing a schematic. This will be more helpful than before.

I notice a couple errors with your schematic, but I don't think they are related to your issue:

This 1KΩ pot (as wired) won't do anything when you adjust it. You're better off using a fixed resistor here if this is how it's wired.

Your recording sounds a lot like one or more of your opamps is running out of headroom, very quickly. How hot is your signal coming into the circuit?

I connected the pins of the unused IC (IC2b) as an independent voltage divider - as suggested by texas instruments. Improvement for sure - no influence on the problem.

You're not forming an independent voltage divider here, you're running it as a non-inverting buffer on your reference voltage that's already buffered.

The buzz noise appeared the first time in a pure breadboard design according to the original single channel design and every test now

I hope you don't take this the wrong way, but I have to ask. Why would you proceed building this circuit without diagnosing the cause of the issue before further complicating matters?

 

[Sonny]

Hi Brett,

thanks for your impressions and questions. Last on at first...

"Why would you proceed building this circuit without diagnosing the cause of the issue before further complicating matters?"

Well, that's a good question and I can't even take it the wrong way , thanks for that. The short and maybe trite answer is:
I was busy with the problem on breadboard for nearly 4-6 weeks and did several of the mentioned tests (in my first post: swapped TI072, 2SK117, run via battery, run out of environement). After failing in diagnose it I draw a wrong conclusion out of the "fact" that this seemed to be a proved unit (according to reports on dirtboxlayout). So I assumed (wrong, as I know now) it had to be some shielding issue based on my experiences with transistor based pedal builds). That's nearly all. At the moment I will try to come as far as possible with actual state...if rebuilding is an option with valid more promise of success - maybe I'd do it as well.

"This 1KΩ pot (as wired) won't do anything when you adjust it. You're better off using a fixed resistor here if this is how it's wired."

The 1k-Pot is an internal trim pot and it does work - meaning it does have an effect on the intensity of the low pass in real - I tested this unit. But indeed the visual is not accurate -I took a symbol for trimpot- in real life one connection is at the wiper for sure.

"You're not forming an independent voltage divider here, you're running it as a non-inverting buffer on your reference voltage that's already buffered."
I apologize for wrong terms here. I did a research cause I needed to know which way I should the unused IC after all. The most valid explanation seemed to be this one from Texas instruments (Link to document). They recommend to connect unused IC in a single supply rail this way



For I already did have a stabilized middle voltage I skipped dividing voltage once more and did connect non-inverting Input to Vbias directly.

"Your recording sounds a lot like one or more of your opamps is running out of headroom, very quickly. How hot is your signal coming into the circuit?"

Thanks to your ears, open. Sorry, I'm not always used to terms, which may natural for you.

What do you mean with the "hotness" auf the signal in particular?

1. The db? Well at the recording the signal comes with extra db from a boost (Lehle sunday driver, set at about noon, so 7-8 db plus) but I recognised the problem first time without a boost in serial singlecoil/humbucker settings of different guitars.
2. Not the db? How and where can I measure it?

Note that the problem noise gets even louder with more gain in the unit:

• Increasing gain on pots (channel normal or bright)
• increasing values in feedback network of the IC (3,3 k feedback resistor up to 12k) as recommended in the project related chat (dirtboxlayout), you see I got trim pots here (I'll add them on the schematic, soon),
  in the first tests on breadboard I was suspicious on the trimpots - but even fixed values didn't change the problem

At last: Does the fact that increasing gain is increasing the problem a clue? In the schematic it seems the gain pots are simply a volume, who decides how much of each channel (after first gain stage) goes into next state. Until now and on base of the audio probe I always thought, the problem would be just amplified in the following stages...

Thanks a lot and greetings,
Sonny