OHHHH I had no idea Leo Graetz is the inventor of the diode bridge ahhh thanks ZM! I learned something else new today!
Back to the problem at hand. What I'm curious about is WHY did this just start happening and why did it only happen to one channel? maybe the bounce has been happening all along and one of the 040's just said I have had enough and died. and the problem is worse with one one amp channel connected to the power supply?
Hope Papa wont mind me posting this. this is the schematic for my 3 Bulb Zen lights as built by N.P. himself (or one of his shop flunkies). this is the schematic for one channel or phase and not show is the relay that shorts the output to ground during power up. the power supply shown is common to both channels.
During power on. The output of my preamp shorts it's output to ground which effectively shorts the input of the amp to ground. this puts C2/C5 in series with R3 to ground this is in parallel with R1 which should bring the gate closer to ground? But im wondering if the C2/C5 combination can be charged up through R2 fast enough to exceed the 20V maximum gate voltage during the brief period while the mosfet is off?
During power on. The output of my preamp shorts it's output to ground which effectively shorts the input of the amp to ground. this puts C2/C5 in series with R3 to ground this is in parallel with R1 which should bring the gate closer to ground? But im wondering if the C2/C5 combination can be charged up through R2 fast enough to exceed the 20V maximum gate voltage during the brief period while the mosfet is off?
OHHHH I had no idea Leo Graetz is the inventor of the diode bridge ahhh thanks ZM! I learned something else new today!
You probably thought : aha the Greats bridge !
I have been doing some reading on thermisters and there seems to be some differing opinions on them. If I use resistors I can do a four step power up sequence using 45, 25, 10 ohm resistors about .5 seconds apart. if i use thermisters I am going to have to order a bunch and play around with the values to get a nice slow ramp up.
heh
over-engineering rings a bell ?
NTC's are having that slow ramp-up already as inherent virtue
I'm always advocating replacing fixed resistors with NTC's in soft start circuits ; almost all of them are always selling as Miracle Elixir from WW , but they really must be optimized for particular xformer/cap bank/load combo
with NTC's , everything is much smoother
over-engineering rings a bell ?
NTC's are having that slow ramp-up already as inherent virtue
I'm always advocating replacing fixed resistors with NTC's in soft start circuits ; almost all of them are always selling as Miracle Elixir from WW , but they really must be optimized for particular xformer/cap bank/load combo
with NTC's , everything is much smoother
It seems my original idea of 45, 25, 10 ohms is a bit much. I found that 20, 10, 5 ohms gave a nicer smooth ramp up with no big jump from 5 to 0 ohms. But I was just experimenting with the light bulbs connected directly to the power supply for a bit over 4 amps total current draw. next i need to connect up some mosfets and see how it reacts.
Zen Mod I will look into some Thermisters now that i have some idea of the resistance values that work!
Zen Mod I will look into some Thermisters now that i have some idea of the resistance values that work!
Welp....Back to the drawing board! I set up a string made up of a zero crossing solid state relay feeding 4x 5 ohm 50 watt resistors with an Arduino micro controller switching 4 relays that would bypass the resistors one at a time 1 second apart and bring up the powers supplies nice and slooow..and for the first half dozen or so tries everything worked fine! then BLAMMO another dead mosfet! this time on the "good" channel!
I need to try disconnecting the speaker shorting relays and try again...I'm running out of mosfets!
I need to try disconnecting the speaker shorting relays and try again...I'm running out of mosfets!
<snip>
I need to try disconnecting the speaker shorting relays and try again...I'm running out of mosfets!
Might need to look no further than here, try series relays instead..
Ah, I didn't realize that there was an output coupling cap in this amplifier.. Series relay clearly is not going to work in this instance.
Looking at the schematic I have a couple of thoughts, one is that this design (obviously based on your experience) is sensitive to how quickly the supply comes up.. I'll wager that something else has changed in your system and there is less IR drop between the wall and your power transformers..
Were it me I would consider introducing a significant time constant in the bias path to the output device, you could do this by splitting R2 and adding a small electrolytic between the resistors and ground. I'd probably make the ratio something like 47K and 9.09K with sufficient capacitance (10uF say) to slow the loop response to 100ms (1tc) and see if that allows you to survive the inrush phase. The ultimate dc operating condition is unchanged, but the loop response which is probably badly under damped should be predictable. Make the time constant longer if you are reproducing very low bass with this amplifier. I believe that the ac performance is dominated by the other feedback loop. Swag on my part.
I'd also add some source degeneration, perhaps 0.1 ohm or so to further improve dc stability..
And add the gate stopper you mentioned in a previous post.
I'm not sure that the feedback components could not be scaled by a factor of 3 or more to get a more benign input impedance at some expense in HF bandwidth due to miller capacitance in the mosfet - what are you driving this with? Even my transformer coupled line stage would struggle with this load..
Yeah, I know this is no longer exactly the zen lite NP designed so feel free to ignore me..
Were it me I would consider introducing a significant time constant in the bias path to the output device, you could do this by splitting R2 and adding a small electrolytic between the resistors and ground. I'd probably make the ratio something like 47K and 9.09K with sufficient capacitance (10uF say) to slow the loop response to 100ms (1tc) and see if that allows you to survive the inrush phase. The ultimate dc operating condition is unchanged, but the loop response which is probably badly under damped should be predictable. Make the time constant longer if you are reproducing very low bass with this amplifier. I believe that the ac performance is dominated by the other feedback loop. Swag on my part.
I'd also add some source degeneration, perhaps 0.1 ohm or so to further improve dc stability..
And add the gate stopper you mentioned in a previous post.
I'm not sure that the feedback components could not be scaled by a factor of 3 or more to get a more benign input impedance at some expense in HF bandwidth due to miller capacitance in the mosfet - what are you driving this with? Even my transformer coupled line stage would struggle with this load..
Yeah, I know this is no longer exactly the zen lite NP designed so feel free to ignore me..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.