LME49990 design note

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
from an LME49990 design note

" SUPPLY BYPASSING
To achieve a low noise and high-speed audio performance,
power supply bypassing is extremely important. Applying
multiple bypass capacitors is highly recommended. From experiment
results, a 10μF tantalum, 2.2μF ceramic, and a
0.47μF ceramic work well.
All bypass capacitors leads should
be very short. The ground leads of capacitors should also be
separated to reduce the inductance to ground. To obtain the
best result, a large ground plane layout technique is recommended
and it was applied in the LME49990 evaluation
board."

In trying the LME49990 with Peter Daniels phono, is quite noisy -- all the HFI/RFI generated by the tonearm wiring are heard. Apparently, power supply bypassing may help.

Does the bolded text actually refer to three caps in parallel for each PS voltage? That's a lot of caps!
 
In this instance it is referring to the bypassing caps placed close to the power pins of the opamps. In other words 6 caps per 8 pin package, 3 caps per polarity, placing the smallest size the closest to the device with next smallest after it etc.

This wont have any effect on anything stray picked up by the tone arm wiring.
 
Member
Joined 2009
Paid Member
It seems like you need some EMC decoupling on the input. A small (NPO) capacitor on the input to ground (close to the input of the op-amp) may be sufficient. But you could also put a ferrite bead in series with the input, before the capacitor to ground.
The power supply decoupling will hopefully keep the op-amp stable, but it will not remove noise coming from the input. With high speed op-amps like this the supply decoupling is very important to keep it stable.
 
It seems like you need some EMC decoupling on the input. A small (NPO) capacitor on the input to ground (close to the input of the op-amp) may be sufficient. But you could also put a ferrite bead in series with the input, before the capacitor to ground.
The power supply decoupling will hopefully keep the op-amp stable, but it will not remove noise coming from the input. With high speed op-amps like this the supply decoupling is very important to keep it stable.

The input impedance of the OPA627/637 is 8 or 9 pF.

You're suggesting another similarly sized cap be used on the input to ground?

As far as I know, the circuit is a clone of the 47 labs phono section, but with different op amps. The LME is different again.

What I find interesting is that the speaker fuse blew. The amp is a phase linear, but the pre has an output coupling cap, so this is not a DC problem -- could there have been over 4A at supersonic frequencies? OR maybe the decoupling caps are bad....
 
Member
Joined 2009
Paid Member
The fact that the speaker fuse blew points to a problem with instability (oscillation). Do you have an oscilloscope to check the output?
One problem with a fast op-amp like the LME49990 is that the supply decoupling is very critical. The LME49990 has a Gain Bandwidth of 110MHz.

How did you build this? On the PCB from Peter Daniel? Have you used the decoupling capacitors recommended by NSC/TI?

First priority is to make sure the op-amps are stable. Then you can deal with the HFI/RFI problems, if they are still present. If they are I would start with e.g. 47pF from the input to ground. If that helps, but perhaps does not remove it completely, I would try to add a ferrite bead (or a resistor of e.g. 100 ohm for initial testing) in series with the input.
 
I'm starting to confuse two seperate tests. There are no supply decoupling caps in the original circuit. I added 1uF tantalum decoupling to the OPA chips. That blew the fuse.
-----------------------

I have two LME49990 smd to DIP adapters. I plugged one straight into the OPA627 socket on the Peter Daniel board. I bought them to replace the 627 on the input opamp. This was as noisy as heck, transmitting all the noise from the tonearm ground. I don't see how PS decoupling will fix that.
 
Last edited:
Member
Joined 2009
Paid Member
I had a suspicion that there was no decoupling in the original circuit. With some luck you may be able to get away with that using a "slow" op-amp like the OPA627 (GBW 16MHz). But with a faster amp like the LME49990 that is asking for trouble.

To get the shortest wires possible your last suggestion might be worth trying.
 
Member
Joined 2009
Paid Member
Which it worth trying.

When you write "This was as noisy as heck, transmitting all the noise from the tonearm ground", what kind of noise are you talking about here? White noise, hum, RFI (from radio stations etc.) or?

What is the impedance of the pick up? If it is not very low you may get more noise with the LME49990 than with the OPA627.
 
I use with lme49990 a combo 100n(NPO smd)/100uF(Silmic). I got no problems with it.

This phono stage (47Labs project) has active riaa correction in the feedback, and HF (2122Hz) are cut in 1 stage. So it should not be a problem. With high gain circuits like this one enclosure is important.
If you have RFI problems cut them on the input.

I got a combo of 47Labs/Phonoclone and I place a RC (2k2/470pF) on the input of first stage. From IN- to ground.

Small ferrit placed on signal wire also can help.

Tantal caps can not cause burning fuse.
 
I would have simply agreed with 5th element, but I note Doug Self's take on power supply bypassing, where he suggests that it's preferable to connect small PSU bypass caps from rail to rail in the case of those applied to opamps and other chips, but to be careful with caps from rail to ground, tending to limit the larger ones to the points where the supply comes onto the board, lest noise is coupled into the ground system resulting in degraded SNR.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.