The curves are there but are called load regulation. Same thing, different view, although not versus frequency. BTW The free TINA simulator from TI has models for these so you can run a sim.
Jan
Last edited:
I think these are tradeoffs in target marketing. The TPS7Axxxx is specifically developed for RF apps in mobile phones so load transients are maybe more important than absolute values.
It is always difficult to makes these comparisons because data sheets are often targeted to a specific market.
Jan
I'm having some issues with a negative version of the 100v super reg I made and I was hoping you could help as I am a bit stumped.
I'm having some sort of strange latching issue on startup.
The lab supply current limiting kicks in once the input voltage reaches about 85v. I tried turning off the current limit and it kills the 82v zener and the source follower attached to it.
If I touch my multimeter probe to the gate of the M2 the lab supply starts spazzing out all over the place with voltage and current glitching everywhere and the zener voltage goes down to about 62v. Strangely if I connect a 10meg resistor between the gate of M2 and ground it does not do this.
I've damaged the opamp twice to the point where it is not completely destroyed but no longer functional in my poking around, this has caused the circuit to function correctly. Since the opamp does not work it swings low and stays low but everything else in the circuit is doing what it is supposed to do under the circumstances, not only that but I've replaced the components on the board multiple times. There is nothing wrong with the parts or the design, it's just some strange phenomena that I can't explain centering around M2 and the zener tripping the current limiting in the supply.
During this latch up M2 gets hot but nothing it is connected to gets hot.
It's very strange I cannot figure it out. I've already built 2 of the positive versions of this circuit and they work just fine.
I'm having some sort of strange latching issue on startup.
The lab supply current limiting kicks in once the input voltage reaches about 85v. I tried turning off the current limit and it kills the 82v zener and the source follower attached to it.
If I touch my multimeter probe to the gate of the M2 the lab supply starts spazzing out all over the place with voltage and current glitching everywhere and the zener voltage goes down to about 62v. Strangely if I connect a 10meg resistor between the gate of M2 and ground it does not do this.
I've damaged the opamp twice to the point where it is not completely destroyed but no longer functional in my poking around, this has caused the circuit to function correctly. Since the opamp does not work it swings low and stays low but everything else in the circuit is doing what it is supposed to do under the circumstances, not only that but I've replaced the components on the board multiple times. There is nothing wrong with the parts or the design, it's just some strange phenomena that I can't explain centering around M2 and the zener tripping the current limiting in the supply.
During this latch up M2 gets hot but nothing it is connected to gets hot.
It's very strange I cannot figure it out. I've already built 2 of the positive versions of this circuit and they work just fine.
You may be exceeding the max Vgs of that source follower. That could then overload the 82V zener. Does the follower have an in-build safety zener on G-S? If not, maybe you can try to put one in.
Edit: isn't that follower the wrong way around? Also noted your zener dropped from 82V to 62V and the max Vgs is 20V...
Jan
Edit: isn't that follower the wrong way around? Also noted your zener dropped from 82V to 62V and the max Vgs is 20V...
Jan
Last edited:
Yeah I thought of that, but during the latch up the Vgs is like 1v. When I touch the gate of M2 with the meter and the lab supply starts spazzing and the zener drops to 62v I don't think the source is not following this voltage although I did not check because the lab supply glitches out of current source mode when it spazzes out and the current draw oscillates all over the map. Not something I want to do for very long.
Also like I said once I break an opamp the circuit works fine. I've thoroughly made sure that it is not the components that are at fault.
Also I did add back to back zeners on M2 to be sure, it didn't help.
This whole thing is weird.
Also like I said once I break an opamp the circuit works fine. I've thoroughly made sure that it is not the components that are at fault.
Also I did add back to back zeners on M2 to be sure, it didn't help.
This whole thing is weird.
It's a negative supply so no. It feeds -84v to the positive side of the opamp and the reference zener acting as a virtual ground. This "latch up" issue seems to begin when the 82v zener starts conducting. I just don't know why. The positive version of this circuit does not have this problem.
The source of M2 follows the supply until the current protection kicks in around the time the 82v zener starts conducting. There's no VGS differential to be worried about here.
Once the overcurrent protection kicks in the entire circuit sits between 81-84v. M2 is still functioning as a proper virtual ground.
I just put some diodes between the inputs of the opamp to ensure they stay within limits, no change in result. The supply still kicks into overcurrent protection mode.
Like I said earlier, if I break the opamp, M2 functions fine, everything functions fine (except the opamp). Only when the opamp works does this happen.
M2 gets hot during the latch up, I don't know where the current is going since the opamp and the resistors attached to it stay cool.
Once the overcurrent protection kicks in the entire circuit sits between 81-84v. M2 is still functioning as a proper virtual ground.
I just put some diodes between the inputs of the opamp to ensure they stay within limits, no change in result. The supply still kicks into overcurrent protection mode.
Like I said earlier, if I break the opamp, M2 functions fine, everything functions fine (except the opamp). Only when the opamp works does this happen.
M2 gets hot during the latch up, I don't know where the current is going since the opamp and the resistors attached to it stay cool.
Can I assume you guys are as clueless as I am about this latch up problem?
I also have another problem. I have a couple of super reg boards I made that a just like the positive reg I posted but without the virtual ground circuitry. It's just a normal super reg with a darlington. I had it up and running @ 1.1 amps no problem, today it is squealing like a dying pig.
I replaced all active components on board but it changed nothing.
The entire circuit is glued in place so literally nothing has changed before it started acting up.
Any idea what could cause this?
I also have another problem. I have a couple of super reg boards I made that a just like the positive reg I posted but without the virtual ground circuitry. It's just a normal super reg with a darlington. I had it up and running @ 1.1 amps no problem, today it is squealing like a dying pig.
I replaced all active components on board but it changed nothing.
The entire circuit is glued in place so literally nothing has changed before it started acting up.
Any idea what could cause this?
Yeah obviously something changed, I was just expressing my point of view. I certainly didn't change anything and I glued the circuit in place to prevent changing variables. I replaced all components on board and it still is oscillating. I was hoping to get some insight into what can cause such an oscillation so I can figure out how to fix it.
Well, please don't take it personally, but why replacing all active parts? Unless you have an idea of what is broken, and replace that, the only result is deterioration of the PCB.
Do you have test equipment? Can you measure things like the DC on a reference diode, probe opamp outputs? If you are going to play with this stuff, there's no way around investing some time trying to understand the circuit and thinking of a strategy for fault finding.
Start by posting the schematic of what you build here. Not just 'it's circuit xyz by Billy Boe over at the other thread'. Nobody is going to hunt for that on his own time.
Another thing to look into is wiring to/from the unit, change in wire dress, different load, additional/changed cap here or there, anything. What is the oscillation frequency, could it be a dimmer or something like that? Move the unit to another room, does it still oscillate? Investigate!
And it's more fun too than just changing everything and hope for the best. ;-)
Jan
Do you have test equipment? Can you measure things like the DC on a reference diode, probe opamp outputs? If you are going to play with this stuff, there's no way around investing some time trying to understand the circuit and thinking of a strategy for fault finding.
Start by posting the schematic of what you build here. Not just 'it's circuit xyz by Billy Boe over at the other thread'. Nobody is going to hunt for that on his own time.
Another thing to look into is wiring to/from the unit, change in wire dress, different load, additional/changed cap here or there, anything. What is the oscillation frequency, could it be a dimmer or something like that? Move the unit to another room, does it still oscillate? Investigate!
And it's more fun too than just changing everything and hope for the best. ;-)
Jan
Last edited:
Because the whole thing functions perfectly fine at DC so I don't know what to replace.
You don't know what your own circuit looks like? This thread is dedicated entirely to that schematic.
I posted a pic of it above. It is not a "dimmer" or something, the opamp itself is literally acting as a very loud and annoying transducer. I wasn't kidding when I said it is squealing like a dying pig.
Last edited:
I think the basic problem is that C10 starts with 0 V but the positive supply pin 7 gets a potential that is more than the non-inverting input => latch-up or even destruction. This level shifter is of a less good design I would say. Despite that the reference ground is quite unstable and also noisy since you have a noisy zener plus a mosfet in series. I would have started to have the Didden design but tuned up the gain instead and let the opamp be driven by 36 V. I have that approach in my pcb's.
The positive version of the reg has no issues though. I omitted C10 and all of the other caps in the negative supply for now too, I didn't want more variables to deal with while I'm trying to figure out the problem.
Normally when all caps are in place the output voltage ramps up in a soft start over a few seconds so no need to worry about C10 in that respect. I'm a bit perplexed how the 80v transistors are surviving the soft start though, I feed it with 110v B+.
The negative version worked fine when I had it on the protoboard.
I've been over every inch of the PCB board and I cannot fathom why it latches up when the positive version does not. The only difference is P channel parts and and opposite polarities.
Technically it doesn't latch up, it just begins to short out somewhere, somehow, once the zener starts conducting, thus causing the lab supply to throttle the voltage.
It is not a defect in the parts or the board, the results are consistent across several different boards I have attempted it on. The positive version works, the negative does not.
I have a cap bypassing the zener, forgot to add that in the schematic. What makes it unstable?
A what?
Last edited:
There are two reasons I suggested you post your circuit:
- You are not just talking to me. You are talking to members of this forum many of them do not know circuits as these by heart, but would be interested to follow this and/or offer insight. That's the nature of this forum. It's not a one-to-one affair.
- You may have made an error adapting a known circuit to your own use. This does happen, and we should always discuss the actual implementation that gives the problem.
Look, people are bending over backwards to help, but you must be forthcoming with the info they need. If this takes longer, people will probably start to ask about pictures of connections and wire dress. All in the interest of solving the issue.
Jan
You have a temp co of -10 mV/Deg C in the Vgs of the mosfet. A bipolar transistor would have -2.2 mV/deg C but I'll guess that type of precision is not your main priority. My gut feeling is that you have not a full grip of what happens during start-up.