PT2399 Latchup

[karbomusic]

I know could be searching and reading since I saw it before but I'm a bit busy and going on the assumption the forum might enjoy the conversation. 

I just built a Madbean Sea Urchin (I etched their PCB which is one of the very rare occassions I build someting of someone else's). Pics are always fun so here it is, built on Saturday and serves as part of my "kick around the house" mini stack sitting on a LM386 amp I designed back in the spring; cool little combo and I'm going to load a few more of my circuits in this style for my home mini jam rig:



I've noticed that it appears to latchup quite easily. I actually just had to plug/unplug the power about 10 times, waiting various amounts of time to allow caps to discharge etc. before I can get it to spin up. I thought it was odd this one does it so much compared to my Echobase which also uses the PT2399. I have a drawer full of PT2399s (well about 10 more) and will swap a few but thought it might be nice to hear some technical mumbo jumbo and opinions about this apparent race condition during power up.

Thoughts?

[anotherjim]

Does the PCB tie pin 3 to 4 (digital ground to analog ground)?
That's a must do in my (small) experience. Looking at the pcb and scheme it would appear it does not.
You can't break it by tying those pins.

Seems to be a difference with newer chips, despite that you probably can measure about 30ohms between pins 3 & 4 inside the chip which is correct. None of mine have worked without a tie, although 10ohm resistance was enough instead of a wired short, I don't find performance worse with a simple wire connection.

That design doesn't over source current from the VCO control pin like some do by allowing the total control resistance to get less than 1k. That's often stated as a cause of latchup.

[karbomusic]

Does the PCB tie pin 3 to 4 (digital ground to analog ground)?
That's a must do in my (small) experience. Looking at the pcb and scheme it would appear it does not.
You can't break it by tying those pins.

Seems to be a difference with newer chips, despite that you probably can measure about 30ohms between pins 3 & 4 inside the chip which is correct. None of mine have worked without a tie, although 10ohm resistance was enough instead of a wired short, I don't find performance worse with a simple wire connection.

That design doesn't over source current from the VCO control pin like some do by allowing the total control resistance to get less than 1k. That's often stated as a cause of latchup.

Thanks Jim....

Just confirming I can/should tie those? As in "can't" above means go for it? 

[anotherjim]

Of course Kary. Can't means should
Although my Dad always told me that Can't means Won't.

[karbomusic]

Thanks Jim!

I wasn't too worried, I'd just grab another out of the box if need be but good to save one if I can. Thanks again.