Drop-in replacement for JRC4558

Jump to Latest
#22
sreten said:
Hi,

I'd hazard a very sensible guess the two chips on the right are
input/output. Replace both with 5532 and the rest with TL072.

rgds, sreten.
Click to expand...

I made that same assumption before popping the PCB out and looking at the traces.

Traces from the line input/output do not seem to go anywhere near either of those opamps.

I am tempted to take hi-res shots of both the front and back of the board, and then create an overlay x-ray view of the PCB layout.
 
#23
If memory serves, JRC 4558s are the real deal, while MC4558s ought to be the "dual 741", i.e. lesser-spec'd parts. I would therefore be guessing that the JRCs do something important, like the main EQ, while some of the MCs may even be used for non-audio functions. This device would have to date from about 1978 to 1983; date codes may be more informative. Layout may well be a bit... oldschool.

It would be a good idea to check supplies, JRC used to make 4558DDs that could take +/-22 V (rather than the more standard +/-18 V). Not like I would expect more than +/-15 in an EQ, but anyway. There must be some voltage regs, too, whose current capabilities and cooling would have to be looked at (not to mention that EQs would not tend to have oversized mains transformers, so input voltage better not be too tight now).
EDIT: Oh, missed that this had already been covered... but do check regulator input voltages. 20 V would be better than 18.
 
Last edited:
#25
The regs should be fine. Two rough and ready quick checks for you.

First measure the voltage going into the regs before you mod the unit. Then measure after the mods. As long as you have sufficient input (>18 volts) they will be fine. 18vac should be giving around 25 volts DC input to them.

Second check, the finger test. Feel how hot they are now and compare again after the mods. If you can hold your finger on them they are fine. Also check those regs for dries while you are at it. Parts like that are favourite candidates for problems like that.
 
#28
I just pulled the lid off, powered up and running with signal passing through.

DC voltage input into the regs is 21.5v. Seems there is considerable loss through the diodes.

The heatsinks are pretty hot. I can hold my finger there, but not for long.

Edit: IR thermometer says ~49C.
 
Last edited:
#30
🙂 lol

OK, 49C is fine. 60-65 is OK (well within the reg specs) but you will start to get local heating of the board and heating of nearby caps.

I'd go with sretens recommend of just using 5532's for the output devices and TL072's for the others. There might be some resistive feed before the regs if you are seeing a measured 18vac on the input to the bridge but only 21 or so on the regs.
 
#31
Mooly said:
Toss up between an LM4562 (much higher power consumption) or a TL072 (great all rounder).
Click to expand...

I think that depends on the circuit. The LM4562 has a much higher bandwidth and would likely oscillate. Adding two small value film caps to the supply pins and ground usually solves it but without this I`m not sure if it would be stable. I did such swap ( NJM4558 replaced by LM4562 in a Clarion car audio CD ) and without the caps it oscillated. The LM4562 may also sound too bright in some circuits if used as a drop in replacement.
 
Last edited:
#33
Mooly said:
Thanks Mario, yes swapping opamps is never totally risk free, in fact I demonstrated that very graphically here,

http://www.diyaudio.com/forums/anal...u-have-checked-see-its-stable-havent-you.html

I'll be honest, I wouldn't envisage major problems with this unit.
Click to expand...
I liked this topic very much!
But in my experience, some times the oscillations make only low H (2H, 3H, 4H...) and the sound becomes "magic"! One time I perceived it in an discrete transistor amp: I tamed the oscillations and the sound becomes "normal, boring"... AH, where is the magic now?😛😀 So we REALLY need an scope or sensitive RF sniffer, or mistakes can occur...
 
#35
Cool discussion, everyone. Some day I hope to learn half as much as some of you know.

With regard to sreten's recommendation and other comments; it does make sense to me that the two genuine JRC4558D are deployed in "key" circuit positions, while the 8 (lower spec. ?) MC4558 chips are deployed in "less critical" positions.

Is this enough that you would all simply "wing it", and replace the JRCs with NE5532 and the MCs with TL072? (despite the fact that I can't confirm for certain what any of these opamps are actually doing in the circuit).

I am poised on the edge of placing a Mouser order. Should I bother with DIP-8 sockets, or just solder direct?
 
#36
Two 5532's for the output stage and TL072's for the rest sounds reasonable. Power consumption is the big concern for me with 10 of them in total.

There is a chance to listen to some different opamps for yourself here,
http://www.diyaudio.com/forums/ever...-trying-understand-what-we-think-we-hear.html

If I would have had enough 4558's I would have included them but I hadn't. The 4560 is the next step up here.

Sockets... I was never a fan but they can have there uses. I'd say yes, fit them.
 
#39
I sat on the fence for many years over the 5532, I suspect because all the ones I replaced were in the output filter and reconstruction stages of CD players where there was significant HF hash present at the opamp inputs. Swapping for FET devices seemed to always improve the subjective result.

Years later I have used the 5532 as a pure analogue gain block and find it very good indeed. Never has an IC divided opinion so much 😉
 
You must log in or register to reply here.
Top Bottom