Author |
Message |
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
|
Back to top
|
|
|
Grumble
Joined: Nov 23, 2015 Posts: 1294 Location: Netherlands
Audio files: 30
|
|
Back to top
|
|
|
wackelpeter
Joined: May 05, 2013 Posts: 461 Location: germany
Audio files: 10
|
|
Back to top
|
|
|
zorius
Joined: May 14, 2019 Posts: 2 Location: hungary
|
Posted: Sun Aug 04, 2019 3:28 am Post subject:
vco range switch |
|
|
I was thinking about something similar, to add to an existing VCO layout (dreadbox Hades). Doepfer designed it this way, with a 5 position switch, but it's not clear how to make it: http://www.doepfer.de/a110_tec.htm
How do You achieve adding those 1V steps? Is it additional to the tune/cv in, so it works paralel with them?
Thanks |
|
Back to top
|
|
|
AlanP
Joined: Mar 11, 2014 Posts: 746 Location: New Zealand
Audio files: 41
|
Posted: Sun Aug 04, 2019 5:00 am Post subject:
|
|
|
For that type of rotary, I usually have some masking tape holding the stopper ring in place (not the low tack stuff!) |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Sat Aug 24, 2019 9:47 am Post subject:
|
|
|
Thanks All, I got the rotary switches from Banggood and beeping them out, they seem fine.
So far the little stopper thingys seem to do the trick so I can constrain it to 5 clicks or whatever.
Next question: how to best create the accurate voltages needed. I am constrained for space so need low parts count.
It's 1V/octave, so I need 5V, 4V, 3V etc.
Easiest/most temp stable I can think of is to use something like a REF02 reference chip and then create a voltage divider from hand-picked 10K 1% resistors. Buffer the rotary switches' output with something like a TL071.
https://www.ti.com/lit/ds/symlink/ref02.pdf
Other suggestions? _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
gabbagabi
Joined: Nov 29, 2008 Posts: 651 Location: Berlin by n8
Audio files: 23
|
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Sun Aug 25, 2019 1:17 am Post subject:
|
|
|
cslammy wrote: | Thanks All, I got the rotary switches from Banggood and beeping them out, they seem fine.
So far the little stopper thingys seem to do the trick so I can constrain it to 5 clicks or whatever.
Next question: how to best create the accurate voltages needed. I am constrained for space so need low parts count.
It's 1V/octave, so I need 5V, 4V, 3V etc.
Easiest/most temp stable I can think of is to use something like a REF02 reference chip and then create a voltage divider from hand-picked 10K 1% resistors. Buffer the rotary switches' output with something like a TL071.
https://www.ti.com/lit/ds/symlink/ref02.pdf
Other suggestions? |
I am using 10V LT1021B as a voltage reference and 10k 0.1% resistor divider but 0.1% is still not enough accuracy for me. Input for octave switch is a voltage follower on OP07C (low cost, low temp drift) but TL071 or so will do.
LT1021-10 has a possibility to set exact 10V by external trimmer (although I didn't used it yet). |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Sun Sep 01, 2019 2:33 pm Post subject:
|
|
|
ixtern wrote: |
I am using 10V LT1021B as a voltage reference and 10k 0.1% resistor divider but 0.1% is still not enough accuracy for me. Input for octave switch is a voltage follower on OP07C (low cost, low temp drift) but TL071 or so will do.
LT1021-10 has a possibility to set exact 10V by external trimmer (although I didn't used it yet). |
Hi Ixtern: I have breadboarded a ref02, with 5x 1K 1% resistors in series. The ref02 seems very accurate without trimming and has good temperature stability.
I am seeing accuracy into the thousands (so, 2V is 2.001xxx, worst case). Question Is that good enough? I am not sure how good this has to be to be audible.
BTW I'd put a buffer (TO071 probably) on the output of the switch. I wasn't planning on buffering what feeds the switch.
another tidbit: the madbeans library 1P12T Eagle device seems to be a match for the banggood 12x1 switch mentioned above. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Mon Sep 02, 2019 4:27 am Post subject:
|
|
|
cslammy wrote: | ixtern wrote: |
I am using 10V LT1021B as a voltage reference and 10k 0.1% resistor divider but 0.1% is still not enough accuracy for me. Input for octave switch is a voltage follower on OP07C (low cost, low temp drift) but TL071 or so will do.
LT1021-10 has a possibility to set exact 10V by external trimmer (although I didn't used it yet). |
Hi Ixtern: I have breadboarded a ref02, with 5x 1K 1% resistors in series. The ref02 seems very accurate without trimming and has good temperature stability.
I am seeing accuracy into the thousands (so, 2V is 2.001xxx, worst case). Question Is that good enough? I am not sure how good this has to be to be audible.
BTW I'd put a buffer (TO071 probably) on the output of the switch. I wasn't planning on buffering what feeds the switch.
another tidbit: the madbeans library 1P12T Eagle device seems to be a match for the banggood 12x1 switch mentioned above. |
1V/Oct means 1000 mV/1200 cents. So every 1mV changes frequency by 1.2 cent. 1% resistors accuracy means that even if you have precise reference voltage (+5V), divider output voltage may differ by 1% = 10 mV to 50mV. And 10 mV corresponds to 12 cents, 20 mV to 24 cents etc. How many Hertz it is? Depends on the VCO frequency. For 100Hz, 10 cents is about 0.58Hz, for 1000Hz it is 5.8 Hz. Using one VCO it may be audible or not (depends on ear) but using two or more VCOs it may be not acceptable.
All depends on your requirements. You may use 1% resistors but select them with greater accuracy (0.1% or better). 1% resistors have usually worse temp drift (50ppm or worse) than 0.1% (25ppm or better) - always check ppm in datasheet.
The most important thing is to have the same values for all resistors, absolute value is not so significant.
TL071 as a buffer: it has low bias current (30pA typ.) but at high temp it may be up to 20nA (worst case). Your divider current is 5V/5kOhm = 1mA.
20*10^-9 / 1*10^-3 = 20/10^6 = 0.00002 = 0.002 %
Looks like input bias current drift doesn't affect accuracy.
Input offset voltage drift: 10uV/deg (typ.) Taking +50 degree temp drift it means that your output voltage will change 500uV = 0.5mV = 0.6 cent.
0.6 cent is not a great value but it adds to all other errors.
OP07C as a buffer: bias current 2.2 nA typ. (14nA max) in all temp range. Input offset voltage drift: 0.5uV/deg. So it is a little better buffer if divider's resistor chain has not too big value (so bias current doesn't affect accuracy) but in your case TL071 should do also.
I am hearing the difference in pitch for a few mV change of control voltage so for me +-1mV accuracy is the desired target. Last edited by ixtern on Mon Sep 02, 2019 10:29 am; edited 1 time in total |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Sat Oct 05, 2019 8:19 am Post subject:
|
|
|
OK, I have laid out my idea for the octave switcher and it doesn't work. Maybe someone has ideas why not?
https://audiodiwhy.blogspot.com/2019/10/cem-vco-part-i-ref02-based-octave.html
If I remove the analog rotary switch the voltages are acceptable. With the switch in circuit it isn't. The switch isn't wired right--it spins the wrong way but that shouldn't affect output V, right?
I can't figure out how the switch could screw up the voltages; it's like it's providing resistance somehow. I was careful soldering it in....trace width? Layout of the board in general? I am a bit stumped!
I am going to dig more into this today, but if anyone has a brainstorm let me know, I am going to need to redo this board to fix the switch and it'd be nice to fix the voltages at output as well! Thanks in advance!!! _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
Posted: Sat Oct 05, 2019 10:24 am Post subject:
|
|
|
Quote: | The switch isn't wired right--it spins the wrong way but that shouldn't affect output V, right? |
Being a single pole switch that shouldn't matter no, and by the looks of it the pins would be correct if you just solder it on the other side of the board.
Quote: | I can assume the rotary switch won't add any resistance (right?) | There will always be some resistance but it is negligible especially compared to the 10K resistors.
Have you tried it without the opamp ? The switch shouldn't cause any problems if you get the correct voltages from the resistor network. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Sat Oct 05, 2019 11:20 am Post subject:
|
|
|
OP07C is soldered wrong on your first photo - pin 3 (non-inverting input) is soldered to C3 but should be to the center pin of the switch.
On the 2nd photo is correct. I don't know which one have you tested...
What exact part numer is for Vishay resistor array?
Redraw your schematic using internal structure of resistor array to see if resistors are connected properly and then check PCB design if resistor connections on PCB are the same as on schematic. |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Sat Oct 05, 2019 1:05 pm Post subject:
|
|
|
Thanks for the suggestions ixtern
ixtern wrote: | OP07C is soldered wrong on your first photo - pin 3 (non-inverting input) is soldered to C3 but should be to the center pin of the switch.
On the 2nd photo is correct. I don't know which one have you tested...
|
Tested the 2nd one. Yes, I figured out that I had wired the OP07 backwards, it was indicated by a dead short in the board when first testing.
ixtern wrote: | Redraw your schematic using internal structure of resistor array to see if resistors are connected properly and then check PCB design if resistor connections on PCB are the same as on schematic. |
The part used is Vishay ACASA1002S1002P1AT
But: redraw--that would mean changing the library part in Eagle, I don't feel like doing that now. It's clear that 1-8, 2-7 etc. are how the part is wired, I can see that with a DVM. What is on the schematic, although not as elegantly drawn as it could be, is correct, again I can confirm with DVM.
OK, I have done more work on this this morning.
I remove the rotary switch all the voltage points are correct. So I know the problem is after that voltage dividers.
If I wire back up the switch with small jumper wire it is unreliable at its output.
I updated the blog post: at least part of the problem is the switch. I should probably try to find one of higher quality, I am pretty sure if you heat up the ones I got from bang good even a tiny bit the guts inside melt. I am having the same issue with 2 of the 10 switches I bought.
I am also going to redo the board with larger traces, mostly because some of them came up while I was using a HAR for troubleshooting. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Sat Oct 05, 2019 1:55 pm Post subject:
|
|
|
Schematic with resistor set drawn this way is rather cryptic so I've asked to show individual resistors but I've checked datasheet by your part number and it seems ok.
Switch may not be OK but also OpAmp voltage follower. Did you checked if it works OK? Connect +5V ref to the +input and check if the same voltage is on output. Also when +input not connected, is there zero voltage on the output (+-offset, some mV or less)?
I've worked with OP07C, have built octave divider with it but DIP8 only version. |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Sat Oct 05, 2019 8:45 pm Post subject:
|
|
|
Gang: I figured it out. Definitely the switch.
I found a 6 x 2 rotary switch in my junk box and wired that in using 22ga. hook up wire. Works like a charm. The voltages are coming in at something like .99982 V, 1.9997 V, 2.999V, 3.9997 etc.
Hopefully good enough for octave switching. I am going to try to hook it up to a VCO tomorrow and see how it sounds.... If the design holds up I'll redo the board with a tougher Switch and I think I'll be good. Thanks again everyone for your help and support. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
Electric Druid
Joined: Mar 13, 2012 Posts: 44 Location: UK
|
Posted: Wed Oct 09, 2019 2:18 pm Post subject:
|
|
|
Rather than a voltage divider to provide different steps, there's alternative approach I saw somewhere using an inverting op-amp. Might have been the Seq. Pro-One, not sure.
Anyway, there's a simple 1V reference voltage, and then the octave switch introduces more equal-valued resistors in parallel with the inverting op-amps input resistor. Hence the resistor value is reduced and gain goes up. Like this:
gain = Rf / Rin
100K / 100K = x1 = 1V
100K / (100K || 100K) = x2 = 2V
100K / (100K || 100K || 100K) = x3 = 3V
etc etc
Don't know if this method provides better accuracy than a divider.
HTH,
Tom _________________ Electric Druid Synth and Pedal DIY website |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Wed Oct 09, 2019 2:42 pm Post subject:
|
|
|
Electric Druid wrote: | Rather than a voltage divider to provide different steps, there's alternative approach I saw somewhere using an inverting op-amp. Might have been the Seq. Pro-One, not sure.
Anyway, there's a simple 1V reference voltage, and then the octave switch introduces more equal-valued resistors in parallel with the inverting op-amps input resistor. Hence the resistor value is reduced and gain goes up. Like this:
gain = Rf / Rin
100K / 100K = x1 = 1V
100K / (100K || 100K) = x2 = 2V
100K / (100K || 100K || 100K) = x3 = 3V
etc etc
Don't know if this method provides better accuracy than a divider.
HTH,
Tom |
This is an interesting approach. I am getting about 3/500V accuracy from my present setup. I don't think I will have to time to check as what I have now is probably good enough but it'd be good to know if what is suggested here yields better results. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
Electric Druid
Joined: Mar 13, 2012 Posts: 44 Location: UK
|
Posted: Wed Oct 09, 2019 4:23 pm Post subject:
|
|
|
I checked - it was the Sequential Pro-One schematics where I saw that.
I'd forgotten it also switches a 100K to ground in one position, so it does 0V, 1V, 2V, 3V. Uses a two-pole 4-way rotary, or one set up like that.
The resistors are 100K 1%, matched to 0.1%, according to the schematic. The -1V reference is just an adjustable divider with an op-amp buffer attached.The -1V adjust is one of the trims you can get at through little holes in the front panel. The op-amps (both the -1V buffer and the octave mixers for both VCOs) are all just sections of an LM348 - not very special. _________________ Electric Druid Synth and Pedal DIY website |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Wed Oct 09, 2019 5:48 pm Post subject:
|
|
|
OK moving forward. With a VCO the REF05 + SMD array sounds good, I have to tweak fine tune on the VCO very slightly now and then after agressively octave switching, but I have to do that on my SH5 at times as well, and that's Roland!
It's good enough for rock and roll. I ordered some robust switches and will redesign the PCB.
I also fixed the symbol layout for the resistor array in Eagle.
btw The arrays are actually pretty great part. It's 4x 10K .1% resistors for about USD .73c each in small quantity from Digikey. It has good temp ratings. Good for a lot of things we do as long as you aren't scared off by SMD. It's an exceptionally tiny part. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
PHOBoS
Joined: Jan 14, 2010 Posts: 5591 Location: Moon Base
Audio files: 705
|
Posted: Thu Oct 10, 2019 7:36 am Post subject:
|
|
|
Quote: | I updated the blog post: at least part of the problem is the switch. I should probably try to find one of higher quality, I am pretty sure if you heat up the ones I got from bang good even a tiny bit the guts inside melt. I am having the same issue with 2 of the 10 switches I bought. |
I would not have expected the switch but I've had similar problems with toggle switches (melting). No problems yet with rotary ones
but sofar they've probably all been alpha switches.
I like the idea of using the resistor arrays, you could use one with 8 so you don't need the seperate resistor.
And yeah they are quite tricky to solder. _________________ "My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Mon Oct 21, 2019 6:48 am Post subject:
|
|
|
[quote="PHOBoS"] Quote: |
I like the idea of using the resistor arrays, you could use one with 8 so you don't need the seperate resistor.
And yeah they are quite tricky to solder. |
Thanks Phobos. Good advice as always, I'll look into 8x resistor arrays, didn't know those existed.
(4x are fun to solder, so 8x must be a real blast )
I need to revise all the VCO boards for the next go around, including the octave switcher, I will build 2 more, but the REF02 board rev1 is in a finished VCO and IMO is working great with a temp front panel; a REF02 chip/resistor array/unity gain op amp, and decent rotary switch is all i seemed to need for this.
Thanks everyone for all the help. _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
|
Back to top
|
|
|
|