New 990 Discrete Opamp - 100% SMD version

[elskardio]

Once you start changing input transistors, and output transistors, heat dissipation capacity, output power, you have a different animal in my opinion.  Those are pretty big changes to the DNA of Deane Jensen's design choices and performance specs.

Nothing will compete against the original through hole design and that was never my intention. If you really need +/- 24VDC into 75ohm, just stick to the original design and please put heatsinks on those poor output transistors.

You can call my SMD version a Mid-Power 990 if you want. But for most application (500 series modules) it will match the performance of the original and the sonic difference is indistinguishable (to my ears and 4 studio colleagues).

 

[JohnRoberts]

I do not wish to encourage this exercise but wakefield makes a heatsink that supposedly works with those devices rated at 55'C/W to ambient.

https://www.mouser.com/datasheet/2/433/217-844229.pdf

Even less enthusiastic, if you really need to drive 75 ohms to +/-24v you could take a page from audio power amp design and use a class G/H output stage topology.

Whatever...

JR

 

[john12ax7]

Being able to drive 75 ohms shouldn't be a requirement.  But you might want to have some current limiting in case someone tries it,,as two 600 ohm loads and a 1:2 transformer gets you there.

 

[JohnRoberts]

Being able to drive 75 ohms shouldn't be a requirement.  But you might want to have some current limiting in case someone tries it,,as two 600 ohm loads and a 1:2 transformer gets you there.

The device dissipation is a longer term effect so perhaps a soft PS rail that is allowed to collapse under high current draw, could support transiently driving peaks at 75 ohm, without releasing smoke.
===
Back last century I made a headphone amp that would happily drive loudspeakers, until the thermal breaker in the wall wart opened up. I made an artificially soft PS (resistor in series with rectifier diode) so under heavy load the PS sagged enough to protect the wall wart when driving loudspeakers.

JR

 

[john12ax7]

What did you use for the 4250A transistors? Doesn't seem to be a SMT equivalent that I could find.

 

[JohnRoberts]

From a PN4250A data sheet . "You can replace the PN4250A with the KSP55, KSP56, MPS751, MPS751G, MPSA55, MPSA55G, MPSA56, MPSA56G, MPSW55, MPSW55G, MPSW56, MPSW56G or ZTX792A"

I am too lazy to check them all, but back decades ago I used lots of  MPSA55/A56 for GP complemantary medium power applications (albeit through hole).

It may be difficult to perfectly match the hfe of the 4250 (is that critical?) but most other specs seem realizable.

I am too lazy to search out a schematic for the 990, but look at how and where the 4250A is used in that circuit for possibly more potential SMD substitutes.

I repeat, I am not encouraging this exercise.... just looking at it from a design engineer's perspective.

JR

 

[john12ax7]

Those seem to be all through hole. Which dataheet are you looking at?

 

[JohnRoberts]

Those seem to be all through hole. Which dataheet are you looking at?

ummm I wrote it was from the (obviously) through hole PN4250....  of course the cross references are TH also but perhaps newer.

You should consider doing a specification search using the 4250A data as a starting point, but since you are unlikely to find a perfect match, you need to look at the schematic and see what the application actually needs.  You can use devices with higher specs, but generally not lower.

JR

 

[john12ax7]

Certainly you can find substitute parts,  but the OP was regarding making an exact clone.  The 2484 is available in SMT, but wasn't aware of the 4250, hence the question.

 

[elskardio]

What did you use for the 4250A transistors? Doesn't seem to be a SMT equivalent that I could find.

I'm using the BC860C with good results.

 

[john12ax7]

I'm using the BC860C with good results.

Thanks

 

[elskardio]

Hi Guys,

here's the latest revision of my SMD 990.
I did a lot of testing to find a low cost alternative to the SSM2212.
I finally found a small SOT363 package made by Nexperia.
The hFE is not as high (the ones I measured where all in the range of 380-410) but It doesn't affect the opamp's performance.
I built 5 so far and DC offset stays in the range of 5-7 mV.
I also made a modification so it can run on +/- 12V to +/- 18V.

Here's a THD comparison between the John Hardy 990C and my CA-990S
(measured at 40dB of Gain (non-inverting config) with a load of 150ohm)

The sonic differences is indistinguishable between both.
Cheers

 

[living sounds]

Looks great!