13600 Linearizing diodes, how do I make them work?

[Scruffie]

I get the principal of their function but in a 'standard' OTA phaser stage such as the ROSS here - http://www.home-wrecker.com/ross_phaser_black.png when I add a resistor to deliver 1mA (i've tried varying currents within limit) from 9V to pins 2 & 5 the chip stops functioning, what am I doing wrong? I can't find any good examples of them in use...

[gritz]

The best info on their use (or rather what thet do and how circuits that use them are different from them wot don't) is in the datasheet itself:

komponenten.es.aau.dk/fileadmin/.../Data_Sheet/Linear/LM13600.pdf‎

Utilising the diodes will mean monkeying with/ adding resistors around the inputs of the OTAs - adding diode current changes the input impedance and effective gain of the OTAs and also tying one input directly to 4V5 whilst the other is attached to 4V5 via a 1k (I think - the image isn't too clear) resistorwill cause a pretty big offset when diode current is fed in - the inputs need to be balanced with equal resistors to vref. In any case it's defo not worth trying to get the LFO section working with linearising siodes - it's complicated.

The short answer is don't bother - the diodes are good at lowering distortion at low signal levels, but guitar fx tend to run LM13x00 inputs pretty hot, where the diodes become less effective. They also do nothing to help noise levels, which we tend to care about more (which is why we run this chip's inputs "hotter). The distortion caused by this OTA isn't particularly objectionable unless the inputs are really overdriven - it's mainly third harmonic.

[garcho]

^ +1

Ray over at MFOS has a short description, too, about why we usually ignore those things but I can't remember where. Dig around his website a little, he's an OTA addict.

[gritz]

...oh, and the diode current gets sunk into the Vref source too, so that needs to be low impedance. Feeding 3mA into a voltage divider made of two 10k resistors will make it swing a long way north of 4V5 (this is something to watch when modding any circuit).

Apologies for the woeful spelling in my previous post btw - I'm not wearing my specs! 

[merlinb]

The diodes are also not very useful. Even if you can get the biasing right, the datasheet claims are optimistic. IME it's not worth the trouble, since with them it's still not good enough for hifi, and without them it's still good enough for guitar!

[Scruffie]

The best info on their use (or rather what thet do and how circuits that use them are different from them wot don't) is in the datasheet itself:

komponenten.es.aau.dk/fileadmin/.../Data_Sheet/Linear/LM13600.pdf‎

Utilising the diodes will mean monkeying with/ adding resistors around the inputs of the OTAs

I'd looked at the datasheet obviously but it wasn't so clear to me, however this paper cleared it up nicely and as you say, I need to add resistors, one mystery solved at least. http://www.nutsvolts.com/uploads/magazine_downloads/11/May%202003%20Ray%20Marston%20-%20Understanding%20And%20Using%20OTA%20OP-Amps.pdf

The diodes are also not very useful. Even if you can get the biasing right, the datasheet claims are optimistic. IME it's not worth the trouble, since with them it's still not good enough for hifi, and without them it's still good enough for guitar!

Hmm... at the moment it's not 'quite' good enough for guitar, I certainly don't want hi-fi but if I really dig in on the bridge pick up the chips distort in a yucky OTA way and I was hoping they might get that last bit out without having to alter other areas.

Back to the drawing board from what you guys are saying I guess!

[Lurco]

"yucky OTA way"? Paul Perry (Frostwave) would not call it that way.

[garcho]

but if I really dig in on the bridge pick up the chips distort

Always? The distortion that renders these OTAs less than hi-fi is tiny compared to something like that, no?

[Scruffie]

but if I really dig in on the bridge pick up the chips distort

Always? The distortion that renders these OTAs less than hi-fi is tiny compared to something like that, no?

My reference buffer may have been set up wrong and causing the distortion  it worked but wasn't doing its job properly and threw the OTA tracking off... oh well, I learnt about OTA linearising diodes in the process!