Hz/volt to 1v/oct

[sahko123]

I'm trying to build a converter which can convert from hz/volt to 1V/oct for a synthesiser and the only thing i could find is the ms-02 circuit below,



The top right circuit being the one i found. I tried to build it out of 4558s but they either got very hot or burnt up. I either need to find a different way to do this or a way to get this to work properly i believe the issue is with input impedance on the 4558s compared to 3140s in the circuit. no matter how its adjusted the output always reads Vcc and nothing else. Have I done a stupid? Is there a better way of getting this conversion accurately?

[SuzyC]

The VCC pin is pin 7 on the CA3140 but on the 4558 power pin is pin 8..that's why the smoke, besides that, there's two op amps in the 8-pin package with the 4558 and no trimming pins.

RTFDS

IMHO this circuit was created by a fool and whoever tries to make it work at more than a narrow range will go tweak-crazy.

Nowadays, a low-cost fast microprocessor with D2A output could make this voltage in to voltage out conversion quite easily.

[sahko123]

I'm aware of the pinout differences I didn't just drop in a replacement this was done on a breadboard. The circuit was constructed correctly just using the elements of a 4558 opamp instead of 2 separate opamps

[SuzyC]

4558's don't get hot unless the output pin is shorted to ground or VCC, else the power supply is connected to the wrong  pins.

In your circuit, the output 4558 positive inp pin is connected to ground, this should force the output to always go negative, unless the input allowed voltage range is exceeded(and you report the opposite to this condition)

When an op-amp is functioning, in all cases, the negative input pin should always (always=within a few mV) equal the positive input pin voltage(in this case 0-volts.).   

The original FET op-amp has a wider input voltage range, and this could also be a problem.

Please RTFSS for the 4558 and the 3140..

Check again your  wiring of all parts, esp. the  power supply wiring.

Is the op-amp inserted backwards? When building the circuit, are you reversing, counting the pin connections backward?

Substituting a 4558 for a 3140 is not going to work where the possibility of exceeding the allowable input voltage ranges of the 4558 are exceeded. A rail<-->rail input voltage range op-amp could be needed here to avoid output inversion(Google output voltage inversion in op-amps.)

In the case of your circuit, when working properly, the + inp of the op-amp will always be at very near ground 0-Volts and safely near the middle of the specified range of allowed input voltages.

[sahko123]

in this case i may just switch to using the 3140 id been putting it off since getting parts in Ireland is a bit of a pain with brexit so iwas hoping to just use what i had for this converter but if in the end i have to fins a similar or equivalent ic i might as well just order the right thing in.

[SuzyC]

I am almost certain your real problem is a wiring error, chip inserted reversal, not due to the substitution of a 4558.

Prove me wrong! 

I deserve to know how much a smarty pants I really am!

Post the voltages measured on each pin and your measured resistance of  (while turned off) each of the 8-pins.

Use your ohmmeter function to measure the resistances at the lowest possible range esp. when reading the output pins of each 4558 sections of the op-amp to the +/- power supply rails and to ground.

Finally I no  know how geeked you are, so I say make sure what you are plugging into the breadboard is a 4558 op-amp (not a 555 timer or something), and that you are counting pins (from a top view of the ic), with the notch at your left with the tiny dot showing pin 1 at the lower left.

 

[Terry Bites]

The 4558 will not substitute for the CA3140. This is because the input bias current of the 4558 is 100nA. This is massive 100000x compared to the 1pA of a CA3140 and will ruin your day. You really do need to trim the offset to get it to work properly. Try that on a 4558! The Korg circuit shown only has to be cover 3.5 octaves. You will have tuning problems on a larger keyboard. https://northcoastsynthesis.com/news/exponential-converters-and-how-they-work/ There's an excellent paper discussing the pros and cons of CV convertors. http://www.openmusiclabs.com/files/expotemp.pdf



A microcontroller with ADC and DAC even PWM>DC will work. Input and output scaling will need to be added.  Watch out, exponential functions which seem ideal, but can take a long time to execute. Lookup tables are easy to generate and area lot quicker.

[Wolfram]

Get the circuit working with the 4558 before replacing it with a CA3140. If you have a wiring error (which seems very likely from your description), you might well toast the 3140 as well, and then you're back where you started.

[sahko123]

1. I'm not sure that anything i said would indicate incorrect wiring seeing as how the circuit was built around the 4558 initially rather than dropping one in where a 3140 would go (and no it most definitely not a 555 timer)

2. The input bias is what i think the problem is and have ordered in ca3140s to use instead. I also noticed that in the same circuit 4558s are used already and it wouldn't make sense why they would use 3140s if 4558s would work. Meaning that the circuit would have been built with the 4558s if it was possible. So instead i will have to trim 3140s to work as needed in the circuit. The CV converter is being made out of necessity and will be used for testing. I have an ms-20 that outputs hz/volt and i need 1v/oct to test a as3340 circuit i have and be able to play with it. I want to use the ms-20's output because i can feed it midi through my computer. So i at most will use around 1-2 octaves anyway. Using a microcontroller isnt really necessary for this reason and its simpler to get a small converter plus i dont have one to hand. If i went the microcontroller route i would just have a midi to 1v/oct anyway rather than bothering to convert it.

3. I have no reason to use the 4558 if i have the 3140 anyway since i tried it because i dont have any 3140s stocked. (but now have them ordered in).

The 4558 didnt actually get toasted just fairly warm. I'll try again sometime later today to see if i can get it to work but again it is most likely that i need the 3140 or equivalent to have it work. Due to both null adjust and having a lower input bias

[sahko123]

to update i have redone the circuit using the CA3140s im after receiving. (I also used two bc337 matched transistors for this rather than the 2sc1583). In terms of functionality im going to let the graph below speak for it self. Yellow being the input sweep from 0.128V to 8V (C-1 to C5). Pink/purple being the desired output (0V-6V linear) and blue being the actual response. (to get the response to start at 0V I have a level shifter that adds adjustable dc offset and adjustable gain). The response needs to be as linear as possible.

This is quite a lot better than the 3.5 octave response because i added a few extra adjustments since the supply is +-12V rather than +12V and -15V of the original circuit