Hello (yay first post!)

I've been looking to find some reliable sources of tempco's mainly 1k 3300ppm used in designs like the Yusynth's VCO.

I found that Farrell are selling two types of LT73 PTC Thermistors:

2k -+5% 3900ppm - https://uk.farnell.com/citec-te-connect ... 317?st=ptc *

2k -+5% 3000ppm - https://uk.farnell.com/citec-te-connect ... 312?st=ptc

I was wondering if I could combine these two types in parallel to create a 1k tempco with a temp coef of 3450ppm? Would this be suitable?

Thanks

My quick maths for finding the temp coef (x):

1/(2000+6T) + 1/(2000+7.8T) = 1/(1000+xT)

Solve for x:

x = (46.8T+13800)/(13.8T+4000)

Therefore for low values of T, x ≈ 3.45 ≈ 3450ppm

* The listing does not give the ppm, but after asking a 'live agent' they told me it had a temp coef of 3900ppm.

## Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

**Moderators:** Kent, luketeaford, lisa, Joe.

### Re: Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

Yes, but you could also use a combination of a 1k 3900ppm thermistor and 2 normal resistors, and it would eliminate 1 of the tempco resistors. 1k tempco in series with an 82 Ohm resistor, then that in parallel with 13.2k, gets you to 3333ppm (the actual ideal tempco) and a 1k net resistance.

https://uk.farnell.com/citec-te-connect ... 15?st=lt73

https://uk.farnell.com/citec-te-connect ... 15?st=lt73

### Re: Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

i agree with jorg, but it is an interesting idea. it will definitely be better than having no thermistor in there, but not as good as an ideal thermistor. unfortunately, its hard to get an ideal thermistor, as they are only specified to +/-10% accuracy on the tempco, so the question is: how far off? the ideal tempco is 3356ppm/C. so if we run some numbers with a 20C rise we get:

3356: 1.0671k

3450: 1.0690k

3900//3000: 1.0689k

so there is a 0.01% error between the parallel resistors and its 3450 approximation, but 0.18% error between that and ideal. since both of these are less than the tolerance of the tempco, id say its definitely worth a try. one caveat here, is that these are just the resistance errors, and that they get compounded in an exponential converter, so the actual oscillator drift will be greater. if you look at figure 8 here:

http://www.openmusiclabs.com/files/expotemp.pdf

you can see that a 3500ppm/C thermistor has 0.5% error at 20C and 5V CV. so id expect the 3450ppm/C to do better than this.

3356: 1.0671k

3450: 1.0690k

3900//3000: 1.0689k

so there is a 0.01% error between the parallel resistors and its 3450 approximation, but 0.18% error between that and ideal. since both of these are less than the tolerance of the tempco, id say its definitely worth a try. one caveat here, is that these are just the resistance errors, and that they get compounded in an exponential converter, so the actual oscillator drift will be greater. if you look at figure 8 here:

http://www.openmusiclabs.com/files/expotemp.pdf

you can see that a 3500ppm/C thermistor has 0.5% error at 20C and 5V CV. so id expect the 3450ppm/C to do better than this.

openmusiclabs.com

### Re: Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

In all these discussions I think back on the decades of music made with vintage synths having 3% or 5% 'tempcos with anywhere from 3000-3900ppm.

I know some ARP clones were successful using 2K tempcos -of a few different ppms- combined with a metal film resistor -something around 28K to get 1k87, too lazy for math right now.

Sometimes I'm not sure if it matters nearly as much as we tend to think it does, when playing with the numbers. Against maybe it does matter more now due to the rise in use of digital and their common use with analog.

I know some ARP clones were successful using 2K tempcos -of a few different ppms- combined with a metal film resistor -something around 28K to get 1k87, too lazy for math right now.

Sometimes I'm not sure if it matters nearly as much as we tend to think it does, when playing with the numbers. Against maybe it does matter more now due to the rise in use of digital and their common use with analog.

**Treat utility modules as stars instead of backup singers.**

**Treat power supplies like Rockstars instead of roadies!**

### Re: Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

And Welcome to MuffWiggler, Volter!

**Treat utility modules as stars instead of backup singers.**

**Treat power supplies like Rockstars instead of roadies!**

### Re: Combining LT73 2k 3900ppm & 2k 3000ppm for Yusynth VCO

Thanks for all the guidance and welcome.

I'll give jorg's suggestion a try, now I just have to design some pcb's.

And yes I agree KSS, I'm planning to add a digital tuning system so a perfect match is not really needed, just preferred.

I'll give jorg's suggestion a try, now I just have to design some pcb's.

And yes I agree KSS, I'm planning to add a digital tuning system so a perfect match is not really needed, just preferred.