The idea behind the device isn't new but this appears to be a more ambitious implementatrion that the pedecessors. I imagine it is basicly an opamp plus voltage and current boost similar to what you find in designs that use an opamp in place of an an input secition squeezed in to a Multiwatt package. I messed around with the opamp as IS a couple of years ago and found it it quite a bit harder than you think it's going to be. Thye part count keeps expanding to the point that the original idea of simplicity starts looking more and more distant.
This isn't ceticsm but appreciation that the device is non-trivial and probly tougher to make work than the LM3886, for instance. Take a look at the application note relating to muting on the LM3886 and then at the one om the LM4702. If something as appearently minor (to the user) as the muting circuit has these extra precautions this suggests to me that there must be some partiularly nasty challenges to be overcome.
Anyway I got the same message to be poatient -- so I'm being patient.
This isn't ceticsm but appreciation that the device is non-trivial and probly tougher to make work than the LM3886, for instance. Take a look at the application note relating to muting on the LM3886 and then at the one om the LM4702. If something as appearently minor (to the user) as the muting circuit has these extra precautions this suggests to me that there must be some partiularly nasty challenges to be overcome.
Anyway I got the same message to be poatient -- so I'm being patient.
Does anyone have one of these built yet? How does it sound? Any tweaks you care to mention? Tricks of the trade that the LM4702 datasheet doesnt mention?
I designed my own 2-sided PCB with 2-pairs of outputs for each channel with expressPCB software. Are MJH11021/2 any good for this application or did I pull one out of my butt? I'd post the PCB but I wouldnt want to be flamed my first couple times on the forum.
BTW: What size are you bypass caps? I have 40,000uF per rail and calculated a sag of 20% off 75V @ 2ohm.
I designed my own 2-sided PCB with 2-pairs of outputs for each channel with expressPCB software. Are MJH11021/2 any good for this application or did I pull one out of my butt? I'd post the PCB but I wouldnt want to be flamed my first couple times on the forum.
BTW: What size are you bypass caps? I have 40,000uF per rail and calculated a sag of 20% off 75V @ 2ohm.
Nice amp driver chip. I'd like to fiddle with it too but I see some obstacles....
1) Hard to justify using the chip primarily for space savings since such an amp will need a huge heatsink (.e.g. +/-70V rails is approx 200W, which in turn means a heatsink approx 0.2degC/W). A big heatsink in turn means more PCB space.
2) Thermal and overload protection does not exist for the output devices. Chipamps have them, many commercial designs have them, but without access to the integrated VAS stage, adding protection might require a bit of creative twiddling with the MUTE pin.
3) Quoted THD is without the output devices. In a blameless Class B amp, a big portion of the distortion comes from the output stage.
Seems to me like this chip is more for customers designing products that need a simple, compact, powerful, decent-performing power amp. I am not saying it's bad or anything but we DIY-ers won't be able to fiddle with tweaks like exotic parts (well, you can still use Vishays externally despite the internal cheap polysilicon resistors), regulated input stage and not forgetting the most popular hi-fi design philosophy - separate PSU for each channel.
OK I'll admit it, what I'm saying is that I find it hard to justify this project economically (gotta say it's cheaper than something available commercially) in order to get project funding from the boss.
PS: It is still a ClassAB amp so perhaps someone might want to ask NSC why the datasheet circuit has no output inductor and/or Zobel?
1) Hard to justify using the chip primarily for space savings since such an amp will need a huge heatsink (.e.g. +/-70V rails is approx 200W, which in turn means a heatsink approx 0.2degC/W). A big heatsink in turn means more PCB space.
2) Thermal and overload protection does not exist for the output devices. Chipamps have them, many commercial designs have them, but without access to the integrated VAS stage, adding protection might require a bit of creative twiddling with the MUTE pin.
3) Quoted THD is without the output devices. In a blameless Class B amp, a big portion of the distortion comes from the output stage.
Seems to me like this chip is more for customers designing products that need a simple, compact, powerful, decent-performing power amp. I am not saying it's bad or anything but we DIY-ers won't be able to fiddle with tweaks like exotic parts (well, you can still use Vishays externally despite the internal cheap polysilicon resistors), regulated input stage and not forgetting the most popular hi-fi design philosophy - separate PSU for each channel.
OK I'll admit it, what I'm saying is that I find it hard to justify this project economically (gotta say it's cheaper than something available commercially) in order to get project funding from the boss.
PS: It is still a ClassAB amp so perhaps someone might want to ask NSC why the datasheet circuit has no output inductor and/or Zobel?
I looked at the datasheet and didn't think the heatsink was THAT big. A board-level HS should be fine. You could always just bolt it to the same HS as the output devices. If one is looking for just 100W amp the HS issue is reduced, anyway.
Right about output protection, but since you mention the b"Blameless", VI circuit from SElf's design would could just be added to the discrete output section. Ditto any speaker protection relays.
If you want simplicity I don't see why with the appropriate gate stoppers, a very simple L-MOSFET output section witjh zener+diode protection wouldn't work.
Right about output protection, but since you mention the b"Blameless", VI circuit from SElf's design would could just be added to the discrete output section. Ditto any speaker protection relays.
If you want simplicity I don't see why with the appropriate gate stoppers, a very simple L-MOSFET output section witjh zener+diode protection wouldn't work.
Sam,
The heatsink I was referring to was for the output devices. With +/-70V rails, it has to be big.
Adding a VI-limiter to the output stage alone may blow the VAS stage if it activates. Hence, VI-limiters are normally used with VAS-limiters.
All I am saying is that while the basic circuit is simple, it doesn't seem to be as robust as a real chipamp.
All in all, just pointing out the limitations here. Hope potential constructors don't feel discouraged. I'd be curious to hear about anyone's DIY attempt.
The heatsink I was referring to was for the output devices. With +/-70V rails, it has to be big.
Adding a VI-limiter to the output stage alone may blow the VAS stage if it activates. Hence, VI-limiters are normally used with VAS-limiters.
All I am saying is that while the basic circuit is simple, it doesn't seem to be as robust as a real chipamp.
All in all, just pointing out the limitations here. Hope potential constructors don't feel discouraged. I'd be curious to hear about anyone's DIY attempt.
Since that's true of any output device with +/-70 rails and not specific to this device, I didn't realize you were writting about the OP devices.
Yup. All the more reson to want to see the insides. Some VAS are supossed to be less suspeptable to this, such a cascoded VAS. I don't know this from personal testing, just mentioning it. This possible would also be a reson to speculate on if you mcan drive L-MOSFETS with this since you can get some protection with a zener + swtich dioo0des between gate and source without (I think) putting the VAS in danger.
Anyway it looks like it will be awile before anyone gets their hands on one of these.
LM4702 max output current is only like +- 5 mA max.
Means the external transistor stage must have high current gain.
This is why darlington transistors are being used in LM4702 datasheet application circuits.
LM4702 is an interesting IC for high voltage.
If you can handle the very limited current output correctly.
For comparison ordinary Op-amps (OPA2134 ...)
can drive output like +- 30-40 mA max.
Means the external transistor stage must have high current gain.
This is why darlington transistors are being used in LM4702 datasheet application circuits.
LM4702 is an interesting IC for high voltage.
If you can handle the very limited current output correctly.
For comparison ordinary Op-amps (OPA2134 ...)
can drive output like +- 30-40 mA max.
Getting back to technical talk. . . .
I think Russ and Metal have hashed out most of the "op-amp" stage electronics, but there has been relatively little talk about the output stage in a while. Since my chips are in the mail according to National, I want to get the rest of the parts together for this amp. I was leaning toward the MJ11032/G as output transistors, both from an availability standpoint and because I haven’t had a ton of experience with different transistors, and they were among the first recommended in this thread (my background is in chemistry, so my electronics experience is with tiny instrumentation signals, not audio output). If there are other suggestions as to what would sound better, I would love to see what others are using/planning to use.
A few questions to Russ or anybody who would be willing to answer: the Qmult transistor from the schematic on the datasheet doesn’t seem to have gotten much discussion, but based on the thermal tracking comments in the data sheet, it seems like this should be pretty similar to the output transistors. Any suggestions as to what transistors would be a good match for the MJ11032’s, or would another set of transistors be a better match?
Also, Re seems to need to be rather beefy, any suggestions on power handling? Back of the envelope says 5W should be sufficient, but I was leaning toward 10W for the sake of headroom. Also, how well should these be matched? My math may be WAY off here, so I would appreciate a second opinion.
I think Russ and Metal have hashed out most of the "op-amp" stage electronics, but there has been relatively little talk about the output stage in a while. Since my chips are in the mail according to National, I want to get the rest of the parts together for this amp. I was leaning toward the MJ11032/G as output transistors, both from an availability standpoint and because I haven’t had a ton of experience with different transistors, and they were among the first recommended in this thread
(my background is in chemistry, so my electronics experience is with tiny instrumentation signals, not audio output). If there are other suggestions as to what would sound better, I would love to see what others are using/planning to use. A few questions to Russ or anybody who would be willing to answer: the Qmult transistor from the schematic on the datasheet doesn’t seem to have gotten much discussion, but based on the thermal tracking comments in the data sheet, it seems like this should be pretty similar to the output transistors. Any suggestions as to what transistors would be a good match for the MJ11032’s, or would another set of transistors be a better match?
Also, Re seems to need to be rather beefy, any suggestions on power handling? Back of the envelope says 5W should be sufficient, but I was leaning toward 10W for the sake of headroom. Also, how well should these be matched? My math may be WAY off here, so I would appreciate a second opinion.
Russ White said:
I think I remember reading that, but I was a little confused about the way you were implementing the output stage. You mentioned you were eliminating a transistor, so I am guessing it was the "Qmult". . .
I guess I will wait and see what you come up with when you have time.
Thanks for the help.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.