[sdiy] low voltage svf with bjts

Chris McDowell declareupdate at gmail.com
Wed Apr 8 19:43:20 CEST 2020


> Cool! Might want to double check how your transistors are biased. Korg 35 style has them biased in reverse saturation mode, so voltage at the base should be higher than the collectors. 


Ah, thank you! I guess I had thought the linear offset in the expo converter, R34 220k in my pic, would be doing the biasing, but it's very clear now why it would not! A resistor to V+ on the bases of Q7 and Q8 has got it working much better.

> Neat!  Do you get much control feed-through at the BP and LP outputs of the 
> filter due to the BE currents of the transistors feeding into the 
> integrators?
> 
yes! haha. it is not terrible but certainly present. reducing the feedback resistor on the first op amp, U2, significantly helps this, while simultaneously severely attenuating the BP output. I don't quite understand why running the whole loop at a low gain, if that is even how one could describe this, helps with feed-through. I did it at first thinking "bjts eh, make signal small why not". It appears that this low gain was helping this thing behave somewhat normal in the absence of any real bias on those transistors. With some sensible bias going, I've changed those resistors back to equal values and am able to massage the rest into working sensibly, I think! 


> On Apr 8, 2020, at 7:35 AM, dave at expeditionelectronics.com wrote:
> 
> Cool! Might want to double check how your transistors are biased. Korg 35 style has them biased in reverse saturation mode, so voltage at the base should be higher than the collectors. 
> 
> On Apr 8, 2020 06:34, Richie Burnett <rburnett at richieburnett.co.uk> wrote:
> Neat!  Do you get much control feed-through at the BP and LP outputs of the 
> filter due to the BE currents of the transistors feeding into the 
> integrators?
> 
> -Richie,
> 
> 
> -----Original Message----- 
> From: Chris McDowell
> Sent: Wednesday, April 08, 2020 9:03 AM
> To: *SYNTH DIY
> Subject: [sdiy] low voltage svf with bjts
> 
> 
> Howdy list,
> 
> I hope every is feeling alright! So I thought why not try to do a state 
> variable with bjts, kind of like the early ms20. One reason why not is that 
> I don't completely understand how to, but I have gotten this working pretty 
> well, and it will simulate down to a supply voltage of 3.3V.
> 
> I took a working filter I've been playing with at 3.3V, and just as a start, 
> almost a joke, just swapped the OTAs for NPNs. the outputs surprisingly 
> still looked like a filter, but this is kind of clearly not right.
> 
> I've left the exponential converter and the two transistors (Q7 and Q8) 
> connected naively, as that's what I have now. What would be the right way to 
> connect these two sections? Q9 is not delivering the ~70uA that it would 
> into an OTA or dummy load, possibly obviously. This technically works but 
> the range is very limited, about a 1:10 change in current / cutoff 
> frequency.
> 
> The point of this I suppose would an easily controllable alternative to 
> twin-t drums, with very low parts count. Plus, JLCPCB.com assembly service 
> is currently out of LM13700s... so I ended up here.
> 
> Cheers,
> Chris
> 
> 
> 
> 
> 
> Virus-free. www.avg.com
> 
> 
> 
> 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20200408/fb812642/attachment.htm>


More information about the Synth-diy mailing list