JE990 Lives!

[rafafredd]

what about the inductors?

 

[Samuel Groner]

Likely to be L1 and L2 (above Q1/Q2), no? :grin:

But the real inductors with 20 uH will not fit here, they are pretty huge.

Samuel

 

[mnats]

Agree totally, Samuel. As we read from Steve Hogan the LM194/394 was Deane Jensen's original inspiration to design the 990.

Looks like a number of Jensen's suggestions regarding layout have been disregarded as well...

So has anyone done a real 990 pcb layout in a 2520 footprint? Seems like a natural for DIY since it is so well-regarded and because the circuit is public domain.

 

[CJ]

JH drops in from time to time.
Past threads have progressed to:

1) Hardy stuff is in-expensive to the point that it is cheaper to buy than to DIY, and

2) the real mojo is in the screening procedure for the parts

since we do not know much about #2, all the more reason to buy it off the shelf.
unless your into the DIY aspect of it.

 

[mnats]

Believe me, I've scoured the threads. I'm totally into the DIY aspect of it, CJ. We'll never know about #2 if no one builds a real one that works.

Maybe the title of this thread should have been JE990 DOA! :sad:

 

[Samuel Groner]

So has anyone done a real 990 pcb layout in a 2520 footprint?

I don't think so. There's the guy who just copied a commercial unit, but that's not my way. I might give it a try in a month or so.

Samuel

 

[mnats]

By copying a commercial unit I guess you mean someone took apart a JH :roll: I think another thread went to hell when his work was disrepected; I don't want any part of it.

The whole reason why I posted to this thread was because I had just finished redrawing the original schematic and had exported a netlist to do a PCB. I figured someone must have done it earlier - it seems such an obvious candidate.

Oh, congratulations on becoming the discrete opamp meta moderator! Someone PMed a link to the current location of the file we've been discussing - could we get it updated in the meta?

http://www.10000cows.com/JE990.pdf

Wish I had seen it earlier - would have saved me a couple of hours trying to figure it all out from the unmarked gerber file...

 

[mnats]

Seems to almost fit - maybe the footprints of the passives will have to shrink - will connect the rest of the dots later - bedtime...

 

[Bauman]

Looking good!! :thumb: And I see the LM in there, cool.

:guinness:
Fabio

 

[Samuel Groner]

Cool! The 100 nF bypass caps could get a 2.5 mm footprint to make things easier, no?

Samuel

 

[tk@halmi]

Nice little board there! I have always wished that the 100nF caps were placed on the host board instead of the opamp. I am not shure if the extra quarter inch distance for this function would make a significant difference.

 

[mnats]

Good eyes there fellas. Those bypass caps were the last components sitting outside the square before I decided to post a draft. I stuck them up in the corner because that was the only piece of real estate left on the board! I like the idea of putting them on the host board or at least shrinking them to 0.1".

So far it seems that if you want to follow Jensen's layout advice and use a 2520 footprint the board will almost certainly have to be double-sided. Oh well, once a prototype is running well we're all just going to have batches of boards manufactured locally from the gerbers, aren't we? I'm assuming of course that you discrete gurus are going to be around when I start running into problems...

 

[Bauman]

Count on me to run some boards and some tests as well!

 

[mnats]

Had a couple of hours for a play today:

Still a bit to do yet but everything seems to fit now...just.

 

[Samuel Groner]

Good work!

The board is 1.1"x1.1", right?

Double-sided is surely a clever idea.

BTW, does anybody know if the original ferrite beads are still available from Jensen? Assembled or just the core?

Samuel

 

[mnats]

The board is 1.1"x1.1", right?

Yes, 1.1" square, not 1.2" like the dale116dot7 design...

Double-sided is surely a clever idea.

Not so clever, just necessary to keep with Jensen's layout recommendations and - let's face it - easier than doing a single-sided board. It will make it possible to do a ground plane too.

The ferrite beads were made by Fair-Rite Products. They don't seem to have the model number listed in the Jensen paper which is 2673225111, but they do have similar looking items: http://www.fair-rite.com/newfair/pr21.htm
Maybe someone in the US can give Fair-Rite or Jensen a call?

Found these locally, which I've been using for the layout:

 

[mnats]

I think it's far enough along to try a home-etch. A couple of things bother me, though: One bypass cap is close-ish to the supply pin and the other is hanging off the opposite extreme end of the long trace. Wish they were both one or the other... The other thing is there is a single wire jumper needed to feed the base of Q8. That's because I wanted to put the transistors near the supply pins as suggested in the AES paper but the rest of the circuit is flipped meaning the positive and negative rails run on opposite sides to the output transistors.

Suggestions? Start over?

Anyway, as suggested in the Jensen paper, the high impedence node is small and shielded, a diode-connected transistor has been substituted for CR3 and oriented so the flat sides can sit belly-to-belly for thermal coupling. CR9 and CR10 will have to go off-board to be thermally linked to the output transistors, also as per the paper. I've stuck both the o/p transistors at the edge so they can be heat sunk to a can, if we're feeling ambitious...

_________________

PRR said: "Any good fool can copy the 990..."

 

[mnats]

Couldn't sleep tonight, so I've flipped the rest of the circuit to follow the positioning of the output transistors - no more wire links needed. Both bypass caps sit closer to the supply pins. Now there are pads near the o/p transistors so CR9 and CR10 can be easily thermally coupled to the transistors without their leads running between components.

Jensen's not answering email about the inductor cores. Rest of the parts are on order...original parts, not subs. Bauman - were you interested in making a batch with the Gerbers?

 

[Samuel Groner]

Thanks for your work--greatly appreciated!

Samuel

 

[Bauman]

Yes man, I can take care of a test batch, contact me at [email protected].

WELL DONE! Looking great really. You took all the care it needs.

Anyway, as suggested in the Jensen paper, the high impedence node is small and shielded

On a side note this is always a good pratice when it comes to DOA layout...

:guinness:
Fabio