Trapezoid VCO

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: lisa, luketeaford, Kent, Joe.

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Sun Jan 26, 2020 2:39 pm

Haralds:Werk wrote:
Tue Jan 21, 2020 9:42 am
I am using a socket and put the tempco above the socket. Socket - tempco - SSM2220. So the tempco is below the SSM but above the socket.
Is it a good idea to put heat transfer compound between the SSM2220 and the Tempco? Or is that overkill?

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Wed Jan 29, 2020 11:48 am

tardishead wrote:
Sun Jan 26, 2020 2:39 pm
Haralds:Werk wrote:
Tue Jan 21, 2020 9:42 am
I am using a socket and put the tempco above the socket. Socket - tempco - SSM2220. So the tempco is below the SSM but above the socket.
Is it a good idea to put heat transfer compound between the SSM2220 and the Tempco? Or is that overkill?
I don't know. I don't have the means to do the measurements in a temperature controlled environment. I won't use transfer compound here. I suspect that I can find it in places where it should not be after a while.
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Wed Feb 05, 2020 8:17 am

Strangely one of the sine shaper circuits does not work as well as the other. Would that be the result of an unsuitable FET? Is it necessary to choose the FETs out of a batch perhaps? Or perhaps badly mismatched 1n4148 diodes? I thought I had used 1n4153 diodes which are better tolerance 1n4148s but I cannot be certain.

Does anybody know what the harmonic spectra of a trapezoid wave is? Like a triangle I guess - rolled off harmonics - but which ones? Even or odd?

I'm loving this VCO.

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Wed Feb 05, 2020 11:22 am

tardishead wrote:
Wed Feb 05, 2020 8:17 am
Strangely one of the sine shaper circuits does not work as well as the other. Would that be the result of an unsuitable FET? Is it necessary to choose the FETs out of a batch perhaps? Or perhaps badly mismatched 1n4148 diodes? I thought I had used 1n4153 diodes which are better tolerance 1n4148s but I cannot be certain.

Does anybody know what the harmonic spectra of a trapezoid wave is? Like a triangle I guess - rolled off harmonics - but which ones? Even or odd?

I'm loving this VCO.
What is the exact problem? Volume? Shape? Can you add pictures of both outputs please. I suspect the BF256B which you used. May be you was lucky with the first one. Try to replace it. I have build this circuit many times in different modules with the given specs with 1N4148 and BF245B without any problems.

The harmonic spectra is described in the original Don Tillmann article part 3. Here is the formula:
You do not have the required permissions to view the files attached to this post.
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Wed Feb 05, 2020 2:58 pm

Haralds:Werk wrote:
Wed Feb 05, 2020 11:22 am
What is the exact problem? Volume? Shape? Can you add pictures of both outputs please. I suspect the BF256B which you used. May be you was lucky with the first one. Try to replace it. I have build this circuit many times in different modules with the given specs with 1N4148 and BF245B without any problems.
The harmonic spectra is described in the original Don Tillmann article part 3. Here is the formula:
Shape - One circuit has a much better shape than the other after trimming for best possible shape. I will try another BF256B.

Errrr - sorry the formula is a bit perplexing to me - I was just wondering generally - does it have both even and odd harmonics like a sawtooth or just odd like a square etc etc?

nigel
Veteran Wiggler
Posts: 679
Joined: Wed Jan 16, 2013 6:49 am
Location: Melbourne, Australia

Re: Trapezoid VCO

Post by nigel » Wed Feb 05, 2020 8:11 pm

tardishead wrote:
Wed Feb 05, 2020 2:58 pm
Errrr - sorry the formula is a bit perplexing to me - I was just wondering generally - does it have both even and odd harmonics like a sawtooth or just odd like a square etc etc?
They're all odd harmonics: the sin() terms use x, 3x, 5x, and so on. (sin() is ... well, sine ... x is the fundamental frequency, 3x is three times that frequency, etc.)
Also, they get small quite quickly - the third harmonic is 1/9 the amplitude of the root (sin(3x) / 9), the fifth is 1/25 the amplitude, etc.

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1502
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: Trapezoid VCO

Post by cygmu » Thu Feb 06, 2020 1:59 am

nigel wrote:
Wed Feb 05, 2020 8:11 pm
tardishead wrote:
Wed Feb 05, 2020 2:58 pm
Errrr - sorry the formula is a bit perplexing to me - I was just wondering generally - does it have both even and odd harmonics like a sawtooth or just odd like a square etc etc?
They're all odd harmonics: the sin() terms use x, 3x, 5x, and so on. (sin() is ... well, sine ... x is the fundamental frequency, 3x is three times that frequency, etc.)
Also, they get small quite quickly - the third harmonic is 1/9 the amplitude of the root (sin(3x) / 9), the fifth is 1/25 the amplitude, etc.
In fact it's the same spectrum as a triangle I think. The only difference is the phase of the harmonics.

Maybe an intuitive way to understand that is the fact that the trapezoid can be created by mixing two triangle waves of the same frequency, but 90 degrees out of phase -- this is shown on Don TIllman's pages of course.

If the ear is really deaf to phase, does this mean a trapezoid would sound just like a triangle?

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Thu Feb 06, 2020 2:38 am

tardishead wrote:
Wed Feb 05, 2020 2:58 pm

Shape - One circuit has a much better shape than the other after trimming for best possible shape. I will try another BF256B.
Compare the input signals of the sine shaper at IC1A pin14 and IC1C pin1. If they are equal try replacing the FET. Are you running it at 15V or 12V? I have no BF256B around, so I can not test it.
tardishead wrote:
Wed Feb 05, 2020 2:58 pm
Errrr - sorry the formula is a bit perplexing to me - I was just wondering generally - does it have both even and odd harmonics like a sawtooth or just odd like a square etc etc?
The trapezoid wave and the triangle wave have the same harmonic structure. Both contain only odd harmonics. The polarities of the harmonics are different. More on this in the original article from Don.
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Thu Feb 06, 2020 4:13 am

Haralds:Werk wrote:
Thu Feb 06, 2020 2:38 am
Compare the input signals of the sine shaper at IC1A pin14 and IC1C pin1. If they are equal try replacing the FET. Are you running it at 15V or 12V? I have no BF256B around, so I can not test it.
Yeh I'll look into that.
This is a really good VCO. The more time you spend with it the more you find. At first I thought the quadrature outputs were only useful for LFO duties but in audio rate - adding and subtracting the waveforms creates a lot of really interesting wave shapes.

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Thu Feb 06, 2020 6:35 am

tardishead wrote:
Thu Feb 06, 2020 4:13 am

Yeh I'll look into that.
This is a really good VCO. The more time you spend with it the more you find. At first I thought the quadrature outputs were only useful for LFO duties but in audio rate - adding and subtracting the waveforms creates a lot of really interesting wave shapes.
I am glad that you like the VCO. If you want to go more experimental disconnect the Waveshaper form the VCO and try different connections. There are some more interesting waveforms in there. I was thinking about do add some sort of patchmatrix or quadrature mixer here.
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Sun Feb 09, 2020 4:15 pm

Haralds:Werk wrote:
Thu Feb 06, 2020 2:38 am
Compare the input signals of the sine shaper at IC1A pin14 and IC1C pin1. If they are equal try replacing the FET. Are you running it at 15V or 12V? I have no BF256B around, so I can not test it.
I tried to trim the second sine shaper again tonight but with no better success. It is still short of creating a nice sine shape.
Yes I have identical triangle waves at pin 1 and 14
I am running the VCO on +/- 15v
Maybe I have been unlucky with the FET. I will buy another

User avatar
guest
Super Deluxe Wiggler
Posts: 4225
Joined: Mon Aug 19, 2013 11:49 am

Re: Trapezoid VCO

Post by guest » Sun Feb 09, 2020 5:10 pm

its possible that the variation in Idss of your JFET is causing it to be off. you want those drain/source resistors to be equal to the zero gate voltage drain to source resistance. you can measure this, and then substitute the matching resistors. take your JFET, and tie the gate to the source. then take your multimeter and measure the resistance from drain to source (put the red probe on the drain). if you want, you can flip it around and measure it the other way and take an average (tie gate to drain).
openmusiclabs.com

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Mon Feb 10, 2020 8:33 am

tardishead wrote:
Sun Feb 09, 2020 4:15 pm
Haralds:Werk wrote:
Thu Feb 06, 2020 2:38 am
Compare the input signals of the sine shaper at IC1A pin14 and IC1C pin1. If they are equal try replacing the FET. Are you running it at 15V or 12V? I have no BF256B around, so I can not test it.
I tried to trim the second sine shaper again tonight but with no better success. It is still short of creating a nice sine shape.
Yes I have identical triangle waves at pin 1 and 14
I am running the VCO on +/- 15v
Maybe I have been unlucky with the FET. I will buy another
Get the specified BF245B. They are still available if you do some search.
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Mon Feb 10, 2020 8:48 am

Haralds:Werk wrote:
Mon Feb 10, 2020 8:33 am
Get the specified BF245B. They are still available if you do some search.
I am using BF245B. I must have a bad one

User avatar
guest
Super Deluxe Wiggler
Posts: 4225
Joined: Mon Aug 19, 2013 11:49 am

Re: Trapezoid VCO

Post by guest » Mon Feb 10, 2020 12:56 pm

the BF245B has a variation in Idss of 6 to 15 mA and a forward transfer admittance of 3.0 to 6.5 mMhos. this means those resistors can span from 150 to 330 Ohms. so its possible that the JFET is fine, but rather that the resistors are off.
openmusiclabs.com

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Mon Feb 10, 2020 2:10 pm

tardishead wrote:
Mon Feb 10, 2020 8:48 am
Haralds:Werk wrote:
Mon Feb 10, 2020 8:33 am
Get the specified BF245B. They are still available if you do some search.
I am using BF245B. I must have a bad one
I thought you still use the BF256B as mentioned in your earlier post. So if you have the right FET check the resistor values. And have a look at the diodes and trimmers. Last resort is swapping the working FET against the not working so you can deduce that the FET is bad..
www.haraldswerk.de
Home of Next Generation Formant

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Mon Feb 10, 2020 2:29 pm

Oh sorry I forgot to mention I did replace the FETs with BF245b
The BF256b FETs behaved like the bad BF245b I have now.
Nearly there but not quite!
Do the diodes have to be matched? Or nearly matched??

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Tue Feb 11, 2020 2:57 am

tardishead wrote:
Mon Feb 10, 2020 2:29 pm
Do the diodes have to be matched? Or nearly matched??
I have never matched them.
www.haraldswerk.de
Home of Next Generation Formant

User avatar
guest
Super Deluxe Wiggler
Posts: 4225
Joined: Mon Aug 19, 2013 11:49 am

Re: Trapezoid VCO

Post by guest » Tue Feb 11, 2020 12:58 pm

the diodes should be pretty close, and typically are. but, if youre seeing a lot of even harmonic distortion, then your diodes or resistors may be off. if youre seeing odd harmonic distortion, then its the JFET/resistor combo.
openmusiclabs.com

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Mon Feb 17, 2020 6:03 pm

Haralds:Werk wrote:
Mon Feb 10, 2020 2:10 pm
I thought you still use the BF256B as mentioned in your earlier post. So if you have the right FET check the resistor values. And have a look at the diodes and trimmers. Last resort is swapping the working FET against the not working so you can deduce that the FET is bad..
Right so I bought a new batch of BF245b from the same supplier (maybe a mistake)
I replaced the FET in the sine shaper that does not work as well with a new FET............no change that I could perceive from the scope.
I checked the resistors etc and theres nothing wrong - you know how it is....... you think you're going crazy - and then I decided as the last resort to swap the FETS.
And sure enough the fault follows the FET. Now the other sine shaper is not as smooth and rounded.
So the fault is indeed with the FET or rather one FET is better suited to the circuit than 2 others I have tried so far.
At this point I am not a bit paranoid of continuing to desolder and replace the FETs for fear of damaging the circuit board traces,
Harald I presume you have never come across this problem before- do you know what supplier you bought your FETs from? If so I would I would rather buy more from his batch. Do you see what I mean?

tardishead
Wiggling with Experience
Posts: 311
Joined: Sat Aug 03, 2013 3:00 pm
Location: London

Re: Trapezoid VCO

Post by tardishead » Mon Feb 17, 2020 6:16 pm

guest wrote:
Mon Feb 10, 2020 12:56 pm
the BF245B has a variation in Idss of 6 to 15 mA and a forward transfer admittance of 3.0 to 6.5 mMhos. this means those resistors can span from 150 to 330 Ohms. so its possible that the JFET is fine, but rather that the resistors are off.
When trimming for best shape I seem to end up taking the "shape" trimmer to the end of its travel to get the best results.
Does this indeed point to the fact that the resistors need to be tweaked?

User avatar
guest
Super Deluxe Wiggler
Posts: 4225
Joined: Mon Aug 19, 2013 11:49 am

Re: Trapezoid VCO

Post by guest » Mon Feb 17, 2020 6:29 pm

first, double check that your trimmer and corresponding resistor to ground (should be 1.5k) are the correct values, and that the 1M resistors are also correct. if thats all good, then try putting some larger value resistors in for the 150ohms. you can try going smaller first, but that is the less likely direction. to go smaller, you can clip some 1k resistors across them and see if helps (you wouldnt have to desolder anything). otherwise, you can lift just one leg of those resistors, and clip in another 150ohms in series and see if it helps.
openmusiclabs.com

User avatar
Haralds:Werk
Wiggling with Experience
Posts: 281
Joined: Fri Sep 09, 2016 3:33 am
Location: Germany
Contact:

Re: Trapezoid VCO

Post by Haralds:Werk » Tue Feb 18, 2020 3:10 am

tardishead wrote:
Mon Feb 17, 2020 6:03 pm
Haralds:Werk wrote:
Mon Feb 10, 2020 2:10 pm
I thought you still use the BF256B as mentioned in your earlier post. So if you have the right FET check the resistor values. And have a look at the diodes and trimmers. Last resort is swapping the working FET against the not working so you can deduce that the FET is bad..
Right so I bought a new batch of BF245b from the same supplier (maybe a mistake)
I replaced the FET in the sine shaper that does not work as well with a new FET............no change that I could perceive from the scope.
I checked the resistors etc and theres nothing wrong - you know how it is....... you think you're going crazy - and then I decided as the last resort to swap the FETS.
And sure enough the fault follows the FET. Now the other sine shaper is not as smooth and rounded.
So the fault is indeed with the FET or rather one FET is better suited to the circuit than 2 others I have tried so far.
At this point I am not a bit paranoid of continuing to desolder and replace the FETs for fear of damaging the circuit board traces,
Harald I presume you have never come across this problem before- do you know what supplier you bought your FETs from? If so I would I would rather buy more from his batch. Do you see what I mean?
No, I never have had any problems with this circuitry. Sorry, I can not recall where I bought them. But I can send you one of mine if you can't get it working with the suggested changes from guest. Looks as if there are fake ones out there.
www.haraldswerk.de
Home of Next Generation Formant

User avatar
appliancide
Super Deluxe Wiggler
Posts: 1475
Joined: Fri Dec 17, 2010 9:51 am
Location: Idiocracy

Re: Trapezoid VCO

Post by appliancide » Thu Feb 20, 2020 2:58 pm

Haralds:Werk wrote:
Thu Feb 06, 2020 6:35 am
tardishead wrote:
Thu Feb 06, 2020 4:13 am

Yeh I'll look into that.
This is a really good VCO. The more time you spend with it the more you find. At first I thought the quadrature outputs were only useful for LFO duties but in audio rate - adding and subtracting the waveforms creates a lot of really interesting wave shapes.
I am glad that you like the VCO. If you want to go more experimental disconnect the Waveshaper form the VCO and try different connections. There are some more interesting waveforms in there. I was thinking about do add some sort of patchmatrix or quadrature mixer here.
Quadrature mixer? Is that another way of saying matrix mixer, or is it different?

socom93
Learning to Wiggle
Posts: 50
Joined: Sun Mar 24, 2019 1:19 pm
Location: France

Re: Trapezoid VCO

Post by socom93 » Mon Apr 13, 2020 2:59 pm

Awesome work !! :yay: :yay: :yay: Thank for the DIY ressources !

Post Reply

Return to “Music Tech DIY”