Bob Pease on the New LM4562

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
P

Portlandmike

Member
Joined 2006
Re: FET's on an NE5534 !

ashok said:
Anyone heard the Fet modified input to the NE5534 ?
Anything at all ?
Unfortunately I cannot find the link to the image that showed how it was done .
I'm sure that will be a very illuminating experiment.
I don't mind making an assembled pcb and shipping it to D.Self or anyone who has access to an AP test set. Anyone ?
Cheers,
Ashok.
Click to expand...


What's this amplifier doing?
Line level?
Phono?

I hate to sound like a advertisement, but the LM4562 is almost a one stop shop. Its bipolar, but its bias current is low because it has bias current cancellation circuitry built in. I've always been biased to bipolar for the sound though. More dynamic and less thin sounding. I only use FETS when you need very low bias current or the source impedance is high.

btw, when using almost any op amp, always try to use low values of feedback resistors when possible. Also try to balance them when feasible.

Best Regards,

Portlandmike
 
ashok

ashok

Member
Joined 2002
"................What's this amplifier doing?......................"

Line level application .
I have only read about it but without any details. So I'm curious . This is not to say that it is better or worse than an LM4562 but just to try it out and "know" what it "really" sounds like. Maybe a piece of crap but at least we really get to know. It's just an expermiment and all experiments are fun.
Cheers.
 
P

Portlandmike

Member
Joined 2006
ashok said:
"................What's this amplifier doing?......................"

Line level application .
I have only read about it but without any details. So I'm curious . This is not to say that it is better or worse than an LM4562 but just to try it out and "know" what it "really" sounds like. Maybe a piece of crap but at least we really get to know. It's just an expermiment and all experiments are fun.
Cheers.
Click to expand...


That's what diy is all about.
Just know that you'll have 10-100x higher offset voltage.
If its just line level, and you don't need gain, and you don't care about offset, just take your two fets (or better yet, bjt's) and do a buffer. load it with a current source biased by a divider between ground and v-. Don't put cap on the gate or base, or do but listen to both.
Besides the sound of a coupling cap, which is best avoided IMO, a discrete buffer as above sounds great! More live sounding then any op amp I've heard.
Spend your money saved from matched xistors on the best coupling cap you can find. And add about 220ohms~680ohms.

Listen. A bjt version of this is beyond the measurement capabilities of a system 1 Audio precision.

Mike
 
ultranalog

ultranalog

Member
Joined 2001
itsonlyaudio.com
wow!

Dear Douglas,

Many thanks for these measurements! Never expected to see you here.

I've wanted a go at these op amps ever since attending Pease's workshop in October.

The common-mode problem is interesting to say the least. I'd be interested in FFTs or sweeps of differential signals with varying common-mode levels. My guess would be the CMFB loop has substantially less GBW than the differential loop in order to keep dissipation low since it drives 4x larger parasitics. It must have been a well-considered trade-off.
 
C

Chucko

Member
Joined 2002
chucko.com/
There's a Linear Technology design note which talks about the effects of non-linear input impedance and source loading:

PDF file

I wonder if this is what's causing the high frequency rise in distortion evident in Mr. Self's non-inverting tests?

Also, I wonder if Mr. Self would be willing to run some video op amps like the LM6172, LT1355, and LT1358 through his tests.
 
P

Portlandmike

Member
Joined 2006
Chucko said:
There's a Linear Technology design note which talks about the effects of non-linear input impedance and source loading:

PDF file

I wonder if this is what's causing the high frequency rise in distortion evident in Mr. Self's non-inverting tests?

Also, I wonder if Mr. Self would be willing to run some video op amps like the LM6172, LT1355, and LT1358 through his tests.
Click to expand...

The LT article is to the point. I've seen the same in ADI or BB, can't remember DS's where it also talked about the importance of matching the inputs impedance!, but that was a fet I think. I'm sure its still important with bipolar inputs though.
Point is to get it low and keep it low. No real reason not to when an op amp has such great drive capability.

I'd rather see the LM4562 done with more like 1k feedback resistors than 10k or 22k.

So for +3 gain something like 1k -in to output, and 1k -in to ground is more like it. If one wants more gain, make -in to ground lower.
 
P

Portlandmike

Member
Joined 2006
tubee said:
This could be a nice shunt regulator for a LM4562. Better then LM78 based ones for sure.

http://www.holgerbarske.com/divfiles/kt506_ccssr_sch.pdf

The LM317 works as 475mA current limit (1.25V voltage loss)

For + and - there must be two of them. It is from: www.holgerbarske.com (netzteiltuning)
Click to expand...


I'd use the fairly new TLVH431 instead of the TL431. Might not even need to bother with the 4562 actually, unless you needed lots of current.

I wouldn't use a IC regulator like a LM317 though. Its a whole control loop mess in itself. If you want really good DC, and short circuit, and thermal protection then IC regs like the LM317 make sense, but for audio.... hard pressed to beat a simple current source, say set from the 431 reference voltage.

Just my opinion of coarse.

Portland Mike
 
F

foo

Member
Joined 2004
Re: Tried LM4562 in my RIAA preamp.

ashok said:
My RIAA preamp has an OPA2134 at the input and followed by an AD826. The equalisation is passive.
Replacing the OPA2134 with the LM4562 softens the sound considerably . Keeping the OPA and replacing the AD826 was more acceptable. They sound very similar but the sound of brass for example seems ( very marginally ) more rounded off ( less bite ) with the LM4562. Bass on the AD826 sounds punchy and on the LM4562 it seems to have a bit more weight though not as punchy.
I tested this over headphones ( Sennheiser HD580). It will have to be done over speakers next time.

Additionally I had an offset problem in my preamp. With the stock chips I got a dc off set less than 1.6mV . With the LM I got around 28mV . I never checked why this was so but the LM will not be used in this circuit. I will now have to try it out as a standard preamp with some gain.
The LM4562 is a good sounding chip .
Cheers.
Click to expand...

Man, You got an osciliating LM4562. Keep the feedback loop to the minimum and the problem is solved.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom