Discrete Opamp Open Design

diyAudio Member RIP
Joined 2005
Scott,
You just brought up a subject that you can obviously answer for me. When building a board and using rosin core solder should the hobbyist clean the boards after assembly? I have never done this and can say that I am not an EE so perhaps I have just been lucky or stupid and have never seen a failure because of flux. If I should indeed clean the boards what is the correct solvent to use? 1,1,1,T is long since not allowed and I wonder what is the best solvent for this purpose. Sorry for going off topic here, but I have often wondered if this is only a problem with the never non-lead based solders or all solders with a flux core. Built a Heathkit clock when I was a kid and it worked for almost 40 years before the digital display started to burn out. What is the failure mode action here with the flux?
In the very old days Freon used to be used. Brad Plunkett mentioned once that at UREI all one had to do to get rid of a grease stain on one's tie from lunch was to walk by the cleaning tank :)

Rosin flux is fairly nonreactive, and unless the board needs to be exceptionally low-leakage it is cleaned off primarily for cosmetic reasons (and as well it can conceal a poor metallurgical joint from visual inspection).
 
Rosin flux is fairly nonreactive, and unless the board needs to be exceptionally low-leakage it is cleaned off primarily for cosmetic reasons (and as well it can conceal a poor metallurgical joint from visual inspection).

Not only. If you manufacture for usage in tropics you have to remember that some molds eat rosin and poop acid. How do I know? I was taught to design and manufacture military equipment for very different conditions.
 
diyAudio Member RIP
Joined 2005
Not only. If you manufacture for usage in tropics you have to remember that some molds eat rosin and poop acid. How do I know? I was taught to design and manufacture military equipment for very different conditions.

I'll use the same excuse I do when I inadvertently use offensive language (I thought Bruce Hofer was going to faint when I said "bull$hit" once, referring to someone's claims about typical noise of switchmode power supplies):

Sorry, I'm from California.
 
I use to purchase freon by the 50 gallon drum back in the day, but alas that is long since over. I think that they would probably put you in jail if you used freon that way today. This was the blowing agent that we used in the manufacture of polyurethane foam horn lenses. Just watch out on a hot summer day, the boiling point is more than low and it would blow up in your face if you were not careful. I suspect that pure Isopropyl alcohol will do the trick. 111T is what we used to use in vapor degreasers before that was banned also. Seems that most of the best solvents have been outlawed in the US but not in countries like China that got a special exemption for many more years to come.
 
I've been busy and missed this, looks like a great idea.

Thanks, Scott. The LTP and VAS are largely your idea, and the output JFET stage is from a posting by John Curl early in this thread. I added the inductor trick, but that's also due to Walt Jung or somebody from AD in the 60's.

I'll just claim credit for cobbling it all together and simulating it in LTSpice with suitable models.
 
One of Brad's posts on base-current shot-noise being an additional contributor to noise seems to have vanished into the ether.

It was well outside my ken anyway, but speaking of noise performance, there seems to be another Toshiba JFET that looks good for both the upper and lower LTP cascode elements - 2sk184 in TO92s. 50V, 15mS and moderate Idss, with good noise performance down to 0.5 to 1 mA with Rg=10 kohm.

The samples I've measured show 1.5 to 2.5 mA Idss, which is low but usable in the LTP.
 
Thanks - great, that settles it. I'll go with J111/J174/J175 (all appear to be multiple-sourced, and I have access to quantity of Vishay/Siliconix J175 locally, which makes the Idss selection easier).

<snip>.

I was just looking at the pdf for the Fairchild devices with a hopeful eye. From what I can see they do not seem to be particularly complementary. The curves appear to be rather different, and iirc the N channel appears to have more gain than the Pch?

Is this going to work? Have I missed something?

http://www.fairchildsemi.com/ds/J1/J175.pdf

http://www.fairchildsemi.com/ds/J1/J111.pdf

_-_-bear
 
... they do not seem to be particularly complementary. The curves appear to be rather different, and iirc the N channel appears to have more gain than the Pch?

Is this going to work? ...

It's only a push-pull source follower, so it will work with maybe asymmetric current swings even with mismatched Yfs, etc. Worst case, it may have 10 dB higher THD than the sim suggests, but it will probably be mostly H2, which won't hurt audible sonics.
 
linux, are you testing overload behaviour? I've only done some skimpy sims but can't seem to stop it from Phase Reversal on +ve overload with diodes in the usual places.

Your circuit has about 15dB more THD than the FET990 or SW-OPA but that's a fair exchange for 3 less active devices.

Won't do 3.162Vp into 15R but that's an unfair headbanging test. :eek:
 
I've only done some skimpy sims but can't seem to stop it from Phase Reversal on ve overload with diodes in the usual places.

Confirmed - the inductor makes the overload behaviour much worse, but it provides higher loop gain in the audio band.

I'll try some of the usual clamps, but it looks like it doesn't like the inputs/output going within 4V or so from the rails. The BJT output stages are much better in this respect.