LM4562, LME49710, LME49720 Start-up Behavior

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
carlalex
Posts: 11
Joined: Sat Sep 19, 2015 6:12 am
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by carlalex »

mediatechnology wrote:I was wondering if a low-power 5V standby voltage-powered op amp could be used as a comparator to detect signal presence and then bootstrap itself and the higher-current balanced input to +15V (and -15V) supply levels.

An input that powers itself down and provides a power on, mute output and/or signal fail indication is kind of neat and an interesting challenge.
Particularly the signal fail/silence sensor part.
That would be cool. Might be outside the scope of my current project. For that, it seems that the consensus is I shouldn't expect major start-up problems with the sorts of stages I am building. Of course, I'll try it on the test bench first.

Thanks. You've been very helpful.

--- Carl
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by mediatechnology »

PS: Any reason you can't use a on/off power switch

What fun is that?
Exactly. It has to be fun.
The NJM2072 is a very cool little part. Looked like it was primarily for vox-controlled recorders.

Going back to the LM4562, LME49720 start-up...

My power-up sequence involved switching a conventional linear power supply on and of.
Electronically switched power sequencing, where the dV/dt might be faster, may produce different results.

One consideration involving bias resistors that I didn't test is the combination of a "too high" bias resistor - one that would cause latch-up in my test - and a big coupling capacitor.
If that big coupling cap has a fairly low impedance on it's left side during power up, then the capacitor charge-up current would also kick-start the bias compensation circuitry.
Another typical situation would be a rail splitter where the voltage divider is bypassed by a big C.
In those cases the bias resistors could be larger than those that latched-up in the test but still work.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by JR. »

carlalex wrote:
JR. wrote:
mediatechnology wrote:I see the NJM2072 has 20K input resistors: http://www.njr.com/semicon/PDF/NJM2072_E.pdf
Unless I read that wrong it's "input resistor" as in just one resistor. I guess you could use two 2072s and hang one off both inputs, but that raises a second issue. The 2072 looks like inverting with gain of 5x and 7V rail voltage so when that 5x gain stage saturates, the input impedance jumps from 20k to 120k (?) at a little over 1V then may get even weirder at over 7v peak.

I repeat my suggestion to drive it from the low Z buffered input or maybe don't use it at all.

JR
I didn't see anything particularly strange when I hit the 2072 with very high signals when I was playing with prefixing gain stages. I was interested in the low-end sensitivity, but hit it with over 20v out of thoroughness since that would be the other effect of a high gain stage in front of it. But, I wasn't looking at it's impedance, only it's behavior. Still, that's well outside it's spec sheet which indicates a limit of Vcc - 1 for Vin, so probably not advised.

--- Carl
I guess I wasn't clear. If you hang it on the very input there is only one of them, which would imbalance the input, but I guess you could add a 20k to ground on the - input leg. Then the input impedance of the 2072 will change when it saturates which can introduce distortion especially when driven from typical hifi source impedance.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by JR. »

carlalex wrote:
Point taken. The making of the thing is interesting in and of itself, as are the things I learn along the way. For what it's worth, the baseline speakers that this project will replace are Infinity Overture 3s. Will I exceed these? Don't know yet. Will I save money? Certainly not. Will I learn? Already have.
===
The NJM2072 crossed my radar reading the ESP articles on soundwest. It's simple enough to use and wasn't hard to source from Mouser. Works like a champ on the bread board, but the input series cap is definitely required even though the spec sheet doesn't really say much about it. (Size that cap with the 20k input in mind - since it seems to form a high pass filter). I'm using it control various SSRs, indicator LEDs, and the 555 to manage the active/standby control pins of various external components.
---
All the PSUs are fused. I'm buiding the assembly as a plate unit on an aluminum plate, but the issue of what the cavity is made from is on my mind. (My last electronics project was an electric brewing system - 25 amps at 240VAC surrounded by liquid - safety was foremost in that project.)
I don't know the model number but hear that Bruno also makes a plate amp version of the hypex amp with a DSP built in , so that would be mission designed and cover the crossover with even more capability.

I've designed analog crossovers and there are things that digital just does better (like delay).

Yup. Current speakers have quite the vocabulary - pops on start-up, a lovely whine if you unplug them. I was mostly asking about serious problems over annoyances.
PS: Any reason you can't use a on/off power switch
What fun is that? Also, it would be a hard sell to the family to manually turn on each speaker one at a time.

Regards,

--- Carl
Outlet srtrip... :lol:

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
carlalex
Posts: 11
Joined: Sat Sep 19, 2015 6:12 am
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by carlalex »

JR. wrote:
carlalex wrote: I didn't see anything particularly strange when I hit the 2072 with very high signals when I was playing with prefixing gain stages. I was interested in the low-end sensitivity, but hit it with over 20v out of thoroughness since that would be the other effect of a high gain stage in front of it. But, I wasn't looking at it's impedance, only it's behavior. Still, that's well outside it's spec sheet which indicates a limit of Vcc - 1 for Vin, so probably not advised.
I guess I wasn't clear. If you hang it on the very input there is only one of them, which would imbalance the input, but I guess you could add a 20k to ground on the - input leg. Then the input impedance of the 2072 will change when it saturates which can introduce distortion especially when driven from typical hifi source impedance.
Yeah. That thought occurred to me when a little later. I'd want to keep it balanced, for sure.

--- Carl
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by mediatechnology »

I agree that the unequal loading with the 2072 looking at one leg is marginal at best.

I wonder if when powered down, the balanced input op amp would produce sufficient AC (steered from input) on it's Vcc, Vee or output pins that could be detected by an AC-coupled always-on NJM2072.
The 2072 would then bootstrap the rails for the balanced input and switch on the following stages.

The 2072 is isolated from the inputs, only one 2072 is needed for a balanced input, and the input stage can also be powered down.
Through the supply pins the external or internal diode thresholds might prevent it from turning on at a low enough level but the 2072 has gain.
For a balanced input there might be enough current through the input and feedback resistors with the op amp off to make it to the output and NJM2072.

Just a thought...
User avatar
carlalex
Posts: 11
Joined: Sat Sep 19, 2015 6:12 am
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by carlalex »

JR. wrote: I don't know the model number but hear that Bruno also makes a plate amp version of the hypex amp with a DSP built in , so that would be mission designed and cover the crossover with even more capability.

I've designed analog crossovers and there are things that digital just does better (like delay).
I think I've seen those. They looked slick. I don't really have a convenient way to drag a digital signal to each speaker. What consumer receivers even have a per channel digital out? A to D back to A may work, and may work quite well in practice, but seems like a hack to work around not having what you would really want - digital straight from receiver.

Also, being a software guy, I've never really liked what appears to be windows only software as an afterthought which seems to be required to configure these. Will I even be able to program it in 10 years? From my Linux machines today?

As for delay, I think I'm passing on trying to deal with that in version 1 of this project. It's a gold plated project, not platinum ;-) I figure if I want to tackle that in the future, it would just come down to a board upgrade. Digital could be done that way too, for that matter. Power amps and the drivers would be the same in these cases.

--- Carl
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by mediatechnology »

viewtopic.php?f=6&t=641&start=25

I just checked this with a powered down THAT1240.
With a high impedance load, low-level signal passes through at essentially unity gain from balanced input pins to output pin.
With a 1K output load, signal is attenuated by about 26 dB - still enough considering the gain of the NJM2072.
I've got a 47 R between the supply pins - below the drop of the ESD diodes signal passes through the part with what appears to be only resistive termination.
As long as it's below the ESD diode forward drop, the surrounding op amp, when powered down, appears to be fairly high impedance.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by JR. »

If the input signal is coming from a HIFI receiver it isn't balanced anyhow. Maybe wire it up impedance balanced but unless you know the source impedance of the receiver that probably isn't worth the trouble. Keeping the signal 0V separate from the shield probably is worth the trouble.

Hanging the 2072 on a differential input may not hurt it all that much.

======

back in the day we used all-pass filters to make short delay's to compensate for driver physical alignment in active crossovers.

--------

have fun... melt some solder, you'll figure out soon enough what is what.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LM4562, LME49710, LME49720 Start-up Behavior

Post by JR. »

It appears that TI has just set an EOL date of late next year on these.

Perhaps more problem calls than sales... :lol:

JR
Cancel the "cancel culture", do not support mob hatred.
Post Reply