I don't. I really have no clue about it. What exactly does happen within this opamp, resulting in swapped inputs?
Best regards!
No wonder, because the inputs are physically fixed and cannot suddenly be re-soldered!
😉
Should I really reveal the secret? I don't like that, it's better to think for yourself - deduction. If nobody has explained the effect here on Monday or Tuesday, I'll be happy to explain the phenomenon, but not immediately.
Now the weekend begins. Internet detox.
HBt.
😉
Should I really reveal the secret? I don't like that, it's better to think for yourself - deduction. If nobody has explained the effect here on Monday or Tuesday, I'll be happy to explain the phenomenon, but not immediately.
Now the weekend begins. Internet detox.
HBt.
Driving the inputs beyond the rails with no current limiting mechanism can result in the input JFETs acting like a parasitic thryristor. It will latch, and then the differential input voltage must be reversed by a large value before it snaps back into normal operation. It will appear to have a reversed phase at the output for part of a cycle until feedback can re-establish itself.
If you stay inside the rails, it is safe and operates normally with low input bias current even if differential voltage is large. For this reason, they make good COMPARATORS, if the 13V/us slew rate is fast enough. If the input current is limited to somewhere around 6-8mA, the actual voltage at the PIN will stay inside the rail. I have verified this driving as much as 160V thru a 47k resistor. The phase won’t swap or cause latch up. Drive it a volt or two above the rail (or below the - rail) directly and it WILL misbehave.
Let me give us a tip:
Two monolytically integrated p channel jFets; i.e. a p-conducting channel is present, which we would like to cut off.
TL07x-series is latch up free.
The thyristor diode is based on the four-layer diode, we see n-p-[n]-Si dioxide-[n]-p-n.
Two monolytically integrated p channel jFets; i.e. a p-conducting channel is present, which we would like to cut off.
TL07x-series is latch up free.
The thyristor diode is based on the four-layer diode, we see n-p-[n]-Si dioxide-[n]-p-n.
“No latch up when common mode range is exceeded”. But not necessarily when supply voltage is exceeded. I think the valid range for common mode input is something like -12 to +15.5, with 15 volt rail. Outside that range, the VAS loses bias and it will no longer work as an amplifier. It will stay well behaved in that nether region, but don’t expect linear amplification.
TL07x or LF35x should not be used for transnova operation. They can not provide enough drive current for EF2/CFP at the power levels typically used. You want a 5532 or something with an equal amount of oomph.
In that configuration latch up can only occur when the boot strapped op amp supply rails collapse. It’s the base current for the drivers that drains the caps. If that current is limited to the 15 mA that a TL071 can supply it probably won’t collapse anyway. But it will clip the amp prematurely if it’s hungrier than that.
In that configuration latch up can only occur when the boot strapped op amp supply rails collapse. It’s the base current for the drivers that drains the caps. If that current is limited to the 15 mA that a TL071 can supply it probably won’t collapse anyway. But it will clip the amp prematurely if it’s hungrier than that.
There is also this from Elector magazine
https://www.elektormagazine.com/magazine/elektor-197907/58401#&gid=1&pid=1
a more powerful currenrdumping version of the above
https://www.elektormagazine.com/magazine/elektor-197907/58401#&gid=1&pid=1
a more powerful currenrdumping version of the above
Here is one way to use output with higher voltages.
There is a voltage gain of 2.1 in the output. Set by R10 and R9.
The output is at +/-23.7 Volt while the supply to opamp is +/-15 Volt.
Power out is 35 Watt.
There is a voltage gain of 2.1 in the output. Set by R10 and R9.
The output is at +/-23.7 Volt while the supply to opamp is +/-15 Volt.
Power out is 35 Watt.
Good old CFP with feedback.
Be easier to thermal track with folded drivers, aka diamond. likely less distortion.
Was fascinated with CFP with feedback, has a trade off of adding little more distortion at high frequency.
likely needs compensation, what is phase margin?
Forgot OPA134 was available in dip8 package/ thru hole
D44,45 transistor pretty linear probably use a pair for better SOA
Opamp of course would have decoupling caps, and shunt regulators as well.
Be easier to thermal track with folded drivers, aka diamond. likely less distortion.
Was fascinated with CFP with feedback, has a trade off of adding little more distortion at high frequency.
likely needs compensation, what is phase margin?
Forgot OPA134 was available in dip8 package/ thru hole
D44,45 transistor pretty linear probably use a pair for better SOA
Opamp of course would have decoupling caps, and shunt regulators as well.
Last edited:
I guess you'd better swap inverting and non inverting opamp inputs, due to the inverting output stage.
Best regards!
CFP with feedback was a fun trick to make up for the 4 to 6 volts you loose using vertical mosfet.
Still a tradeoff with high frequency distortion, fun topology to play with regardless.
Still a tradeoff with high frequency distortion, fun topology to play with regardless.
Here on forum I was reading of 2 very interesting approaches, big differences and big similarities, but definitely different, kudos to inventors:
So I've got round to testing the physical embodiment of this NIC based amp of mine (unloaded), Approx +/-45V supplies, having now got a QA403 audio analyzer.
First distortion v. input level plot at 1kHz:
Which seems respectable - whether it can maintain this under load has to wait for me to dig out the bigger PSU and dummy load to test.
And distortion v. frequency at -5dBV input level:
The glitch at the end is the third harmonic going out of range - this mode of measurement seems not to allow setting a measurement BW.
Clearly distortion would be rising from...
First distortion v. input level plot at 1kHz:
Which seems respectable - whether it can maintain this under load has to wait for me to dig out the bigger PSU and dummy load to test.
And distortion v. frequency at -5dBV input level:
The glitch at the end is the third harmonic going out of range - this mode of measurement seems not to allow setting a measurement BW.
Clearly distortion would be rising from...
Hi!
I am fiddling around with a couple of OP-input power amps. Wonder if anyone has comments on these:
Magnat Crusader 720 and Montarbo 1002
My schematics are made as driver boards, requiring axternal OPS. Also I have used dual OP so that I can include a buffered volume control.
Ignore supply voltage values, the may be different.
Magnat original:
Magnat my take:
Montarbo original:
Montarbo my take:
I am fiddling around with a couple of OP-input power amps. Wonder if anyone has comments on these:
Magnat Crusader 720 and Montarbo 1002
My schematics are made as driver boards, requiring axternal OPS. Also I have used dual OP so that I can include a buffered volume control.
Ignore supply voltage values, the may be different.
Magnat original:
Magnat my take:
Montarbo original:
Montarbo my take:
Why not just use a high voltage opamp. I measured a OPA454 with+ - 37v rails and the distortion at 10v RMS out into a 10k load was negligible.
Im going to test it as the VAS/input stage for my EL34 SE amp.
There are higher voltage and lower noise versionsn: OPS456 etc
Im going to test it as the VAS/input stage for my EL34 SE amp.
There are higher voltage and lower noise versionsn: OPS456 etc
I tend to stay away from complicated solution, considering it is not hard to add a long tail pair input stage at the front instead of opamp.
High voltage opamp is a good solution.
High voltage opamp is a good solution.
This variation on CFP with gain avoids the high-power feedback resistors, gets rail-rail output and good THD with minimal idle current.
There is also a natural current limit which you can set with R17 and R10.
There is also a natural current limit which you can set with R17 and R10.