Author Topic: transistor replacement for VCO?  (Read 10906 times)

0 Members and 1 Guest are viewing this topic.

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
transistor replacement for VCO?
« on: June 21, 2017, 08:16:38 am »
I would like to experiment a bit with a VCO, as described on page 13/14, figure 14/15 in this document: http://www.ti.com/lit/an/snoa665c/snoa665c.pdf
I can get all the OpAmps, comparators and voltage references, but the transistors are out of production or really expensive. Which replacement parts could I use? I need the following transistors:

- 2N3810
- LM394
- 2N4392
- 2N2369

Would be good to get it all at Digikey. Some of the transistors are available there, but would be nice to have everything as SMD, too and not too expensive (usually transistors should cost only cents). And is there a SMD alternative for the LM329?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline JPortici

  • Super Contributor
  • ***
  • Posts: 3573
  • Country: it
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #2 on: June 21, 2017, 11:28:14 am »
AS394: http://www.ericasynths.lv/en/shop/diy/diy-accessories/as394-matched-transistors/
:) works well

For all of them? And EUR 3 is still expensive, and I don't get them as SMD at Digikey. But I found another IC, the LM13700. Looks like it is easy to create a VCO with it, and it is available for not too much money in SMD at Digikey :)
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline JPortici

  • Super Contributor
  • ***
  • Posts: 3573
  • Country: it
Re: transistor replacement for VCO?
« Reply #3 on: June 21, 2017, 12:35:48 pm »
yes, it is. you have plenty of examples in the datasheet

there are COUNTLESS vcos build with the ca3080 and its derivates (cem chips, lm13700 and such)

there is also a tri->sine waveshaper which uses the arctan transfer function of the inputs, but i don't remember if it's included in the datasheet
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Re: transistor replacement for VCO?
« Reply #4 on: June 21, 2017, 05:35:50 pm »
Curtis CEM3340 VCO is back in production, popular for analog synths since the 1980's.

Linear Systems makes several dual matched transistors, dual JFETS too. Example LS311 dual monolithic NPN transistors.
« Last Edit: June 21, 2017, 05:37:34 pm by floobydust »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: transistor replacement for VCO?
« Reply #5 on: June 21, 2017, 06:51:59 pm »
For the 2N4392, there are direct SMD replacements, like MMBF4392. Much easier to get and less expensive. The two 2N3810 don't look critical at all - so could be just simple PNPs (e.g. 2N3906 like) or maybe a BCV62  (PNP pair for current mirrors).
The LM394 could be replaced with a differential pair in a different circuit. These tend to be still expensive. For a first test one could use cheaper dual transistors - there are a few cheap ones in a SOT23-6 like case.

The 2N2369 could be a little tricky, as it seems to be used for it's short time on saturation. One could be bold an replace it with a small MOSFET like 2N7000 and adjust the resistors. Another option might be a more normal transistor with a schottky diode to avoid saturation.
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #6 on: June 21, 2017, 08:12:07 pm »
there is also a tri->sine waveshaper which uses the arctan transfer function of the inputs, but i don't remember if it's included in the datasheet

Thanks, I found one here:

https://wiki.analog.com/university/courses/electronics/electronics-lab-12sg

It is using a SSM2212 matched NPN transistor pair, which is quite expensive, but the match is 0.5%, compared to like 10% for cheaper pairs. I guess this will influence distortion a lot? But this looks useful, because it is easier to generate a nice triangle wave and then just convert it to a nice sine wave. And might be interesting to feed it with other signals than triangle as well.

For the 2N4392, there are direct SMD replacements, like MMBF4392. Much easier to get and less expensive. The two 2N3810 don't look critical at all - so could be just simple PNPs (e.g. 2N3906 like) or maybe a BCV62  (PNP pair for current mirrors).
The LM394 could be replaced with a differential pair in a different circuit. These tend to be still expensive. For a first test one could use cheaper dual transistors - there are a few cheap ones in a SOT23-6 like case.

I guess I have to order a few different type and try it. Do you have any recommendations, maybe cheaper than the SSM2212, but not too worse, like better than 10% matched gain?

Quote
The 2N2369 could be a little tricky, as it seems to be used for it's short time on saturation. One could be bold an replace it with a small MOSFET like 2N7000 and adjust the resistors. Another option might be a more normal transistor with a schottky diode to avoid saturation.

Where do I need to place the schottky diode and which transistor would be normal at this place?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: transistor replacement for VCO?
« Reply #7 on: June 22, 2017, 06:38:52 pm »
Transistor matching is not that critical. The main error is due to offset in VBE, and there is an adjustment for this anyway. So for a first test one could use the cheap BC847S or similar. I don't think the difference will be so large, that one really needs a much better one.

Temperature stability is a problem with this circuit anyway. So if higher quality / accuracy is wanted the next step might be temperature stabilization or maybe compensation (e.g. use a PT100 instead of the 200 Ohms from one base of the transistor pair to GND).

One could also use an other chip instead of the transistor pair. A MC1496 would be one option if there is sufficient supply. I know it is a mixer but can also be used as a differential amplifier with cascode of the output side.

To speed up a transistor by preventing saturation one can use a small schotky diode from collector to base, just like it is used in the 74LSxx gates.

As an alternative one could also replace the JFET switches and the transistor with a CMOS switch like a 74HC4053. However it needs a suitable supply (e.g. +-5 V), which might not be good for the LF356.
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #8 on: July 12, 2017, 06:45:23 pm »
Ok, I got all ICs for the VCO, and a bunch of transistors for testing the replacement parts. This is the circuit I'm trying to build, from the application note in my first post:



I downloaded all datasheets for the packages and copied it to an image, for easier reference (always a good idea when building it on a breadboard)



Sorry for some noob questions, but how do I connect the LM311? I guess the "EMIT OUT" pin has to be connected to -15V, and "COL OUT" is the op-amp output pin in the circuit? And Vcc+ is connected to +15V, and Vcc- to -15V. But what do I do with BAL/STRB and BALANCE?

And for the LF351 / LF356 op-amps, do I need to connect the offset null potentiometers, or can I leave it unconnected?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #9 on: July 12, 2017, 11:44:47 pm »
Ok, I connected the LM311 as I wrote in my last post. Then I used a 2N7000 instead of the 2N2369, and 22 nF instead of the 10 nF for the integrator (because I want lower frequencies than 30 kHz max). For the 2N7000 I simply used a 1k resistor instead of the 10 k resistor to the gate, and removed the 5 k resistor. And I added 100 nF decoupling capacitors for each op-amp power supply. The rest is the same. I can adjust from about 400 Hz to 10 kHz, not down to 1 Hz as written in the application note. And for higher frequencies it looks distorted:



For lower frequencies, it is less distorted, but not very symmetric:



Green trace is the gate of the 2N7000, yellow trace the "output to log shaper-stage" point.

This is my setup:



The first op-amp on the left is just a voltage follower for the control voltage input. Is it possible that the setup causes the distortion and I should create a proper PCB for it, with big ground plane etc.? And why doesn't it go down to 1 Hz?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline RandallMcRee

  • Frequent Contributor
  • **
  • Posts: 544
  • Country: us
Re: transistor replacement for VCO?
« Reply #10 on: July 13, 2017, 12:24:51 am »
Have you seen the analog parts kit?

http://store.digilentinc.com/analog-parts-kit-by-analog-devices-companion-parts-kit-for-the-analog-discovery/

It contains the parts (e.g. the SSM2212) needed to construct the projects shown in the analog wiki that you linked to. All the parts are breadboardable even the surface mount ones, using adapters. Very reasonably priced, I think.

There are also some pretty cool things like sensors, etc. You probably don't need the resistors and capacitors but it seems to me the ICs alone are worth the price.
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #11 on: July 13, 2017, 05:31:29 am »
Thanks, interesting selection of parts in this kit, but did you add the individual parts cost? Might be the same when buying it at Digikey, and you are right, I have already kits for capacitors and resistors.
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #12 on: July 13, 2017, 12:03:50 pm »
Forgot to mention: I also changed the 2N4392 with the easier to get MMBF4392. And adding or removing the 15 pF capacitor near the LM311 doesn't change anything.
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #13 on: July 14, 2017, 06:46:54 pm »
I found the problem: I used 10 pF instead of 10 nF for the integrator capacitor. Looks good now, with 20 nF for lower maximum frequency:



So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #14 on: July 15, 2017, 12:58:02 pm »
Full working circuit:



I replaced the LM329 with a TL431, because the LM329 is not available as SMD. And I added a frequency range switch: Now I can adjust from about 0.1 Hz to 100 Hz, and from about 100 Hz to 20 kHz.

But the main modification is two inputs for control voltages: With one input you can select the rising speed and with the other the falling speed of the triangle. With this method I can adjust it continuously from nearly sawtooth, to triangle and to inverse sawtooth. The control voltage will be created by a DAC from a microcontroller, so no problem to do some calculations, to make the user interface easier, e.g. the user selects just the frequency and the angle, or pulse width for the square output.
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline orolo

  • Frequent Contributor
  • **
  • Posts: 352
  • Country: es
Re: transistor replacement for VCO?
« Reply #15 on: July 15, 2017, 02:29:09 pm »
The 2n2369 can be found in SOT-23 format for 0,056 euro each if you buy 100. I got from mouser a hundred recently. Really handy transistor for fast switching.
« Last Edit: July 15, 2017, 02:32:55 pm by orolo »
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: transistor replacement for VCO?
« Reply #16 on: July 15, 2017, 05:11:24 pm »
I found the problem: I used 10 pF instead of 10 nF for the integrator capacitor. Looks good now, with 20 nF for lower maximum frequency:





No, it does indeed look bad. Guessing from the ugly distorted triangle sides, you have used a pisspoor ceramic cap for the integrator. that is a big NO! Never ever use pisspoor ceramic caps in any signal circuits. Use only foil type (any foil, is way better than ceramic). Ceramic are good only up to some nF when C0G / NP0 dielectric material is chosen. All other dielectric types are extremely voltage depndent and non-linear.
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #17 on: July 15, 2017, 06:37:40 pm »
And finally the triangle-to-sinus wave-shaper circuit. The replacement parts work well, I just had to change a few resistors:



Examples:





Interesting: if I adjust the wave-shaper pot, I can smoothly change between sinus and square, or sinus and triangle:



Full complete KiCAD circuit:

http://www.frank-buss.de/tmp/vco.zip
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #18 on: July 15, 2017, 06:41:14 pm »
No, it does indeed look bad. Guessing from the ugly distorted triangle sides, you have used a pisspoor ceramic cap for the integrator. that is a big NO! Never ever use pisspoor ceramic caps in any signal circuits. Use only foil type (any foil, is way better than ceramic). Ceramic are good only up to some nF when C0G / NP0 dielectric material is chosen. All other dielectric types are extremely voltage depndent and non-linear.

Thanks, you are right, the curve could look cleaner. I will try to use foil capacitors. Don't have them in my parts bin, but I will order some for testing.

Looks like Digikey has no foil capacitors as SMD (except one part)? Are there only through-hole foil capacitors?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline RandallMcRee

  • Frequent Contributor
  • **
  • Posts: 544
  • Country: us
Re: transistor replacement for VCO?
« Reply #19 on: July 21, 2017, 01:36:03 am »
Digikey says 3,000+ film capacitors surface mount.

Also, one of the really nice things about the Analog parts kit is that you cannot buy the individual components without having to, in many cases, also buy an adaptor and have enough skill to solder them. So, for me, a real bargain. Soldering SMT is not my strong suit.

 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #20 on: July 21, 2017, 06:16:11 am »
Digikey says 3,000+ film capacitors surface mount.

Right, I guess I clicked a wrong search filter. Which capacitor would be good for the integration capacitor? I guess the main problem with ceramic capacitors is the non-linearity, because they change capacity with voltage, and with temperature. I've read that polyester is much better. Would this capacitor be a good choice: https://www.digikey.com/short/3vmq7t  ?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: transistor replacement for VCO?
« Reply #21 on: July 21, 2017, 08:15:27 am »
Nice circuit. The problem I found was the comparator's response time limited the high frequency performance and could be responsible for the distortion at 30kHz.

The sine wave isn't great though. Have you tried a different wave shaping circuit? Here's one I've used in the past. Unfortunately I can't find the original source of the schematic.

 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: transistor replacement for VCO?
« Reply #22 on: July 21, 2017, 11:54:39 am »
Usually the saturated differential amplifier gives a reasonable good sine, if the amplitude is adjusted correct. This might need to have a temperature dependent resistor. Like with the other sine shapers it needs a reasonable good adjustment.

For this relatively simple circuit about any film capacitor is OK. So in this case polyester type is good enough, though polypropylene (PP) and polystyrene (PS) caps would be high quality. But one would not notice the difference.
There are also SMD film caps, but they are somewhat sensitive to temperature during soldering and thus a special relatively temperature stable type. So they are OK though somewhat expensive and more difficult to solder. For the smaller caps NP0 ceramics is a good alternative in SMD. Modern MLCC ceramics of NP0 type are even available in the 100 nF range. However large form factor ceramics are somewhat sensitive to board bending. Als an example :
https://www.digikey.com/product-detail/en/murata-electronics-north-america/GRM31C5C1E104JA01L/490-1767-2-ND/586653

So if would be only the µF range where you would really need a THT film cap. 
 

Offline FrankBussTopic starter

  • Supporter
  • ****
  • Posts: 2375
  • Country: de
    • Frank Buss
Re: transistor replacement for VCO?
« Reply #23 on: September 30, 2018, 07:41:25 am »
Curtis CEM3340 VCO is back in production, popular for analog synths since the 1980's.

I tried to build an oscillator with the CEM3340, but there is a problem with the sawtooth output: after half of the output wave, there is a high frequency wave on top of it, about 3.6 MHz, see this log entry:

https://hackaday.io/project/160538-cem3340-module/log/151718-first-revision-tested

Today I soldered the second version, with all the changes I described at the end of the log and what you can see in the circuit diagram, and it has the same problem, so I guess it is not related to the layout. I even unsoldered the 1 nF Mica cap and connected it directly to pin 11 and 12, but same problem. Maybe there is a problem with the trimmers, or do I need to change some part values?

Circuit diagram:

https://cdn.hackaday.io/images/1641441535289908627.png

What could cause this strange high frequency oscillation?
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: transistor replacement for VCO?
« Reply #24 on: September 30, 2018, 11:15:58 am »
What could cause this strange high frequency oscillation?

LF356 wave-shaper circuit: amplitude POT with wires may convert intended gain stage to generator. CEM3340 circuit may also have feedback somewhere. To suffer all you need is fast opamp on breadbord and some wire ;) Solution: replace POTs with SMD resistors placed close as possible to opamps and see if it helps.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf