Eva said:
Hi Eva, nice to hear from you.
OK, running in linear mode makes sense - and I'll try that in simulation. It's got to be a bit tricky to run linear, though because of the rotten tolerance on current transfrer ratio. For example, the CNY17-1 has transfer of 0.4 < CTR < 0.8. Some of the other versions have higher transfer ratios. It's really easy to make these devices into switches, but running linear can't be easy. Say I buy a bit lot of them, who's to say what CTR I get? Does every circuit built have to be manually adjusted? They can't be run with feedback because of the isolation barrier, but maybe another method is to run 2 optos, making a nice feedback loop. Is anyone doing such a thing? If not, can you point me to some examples?
gene
gearheadgene:
It's actually quite easy to ensure linear operation, the bias point is adjusted through the main feedback loop.
Check out this schematic, it's a basic arrangement and it's quite self explaining:
It's actually quite easy to ensure linear operation, the bias point is adjusted through the main feedback loop.
Check out this schematic, it's a basic arrangement and it's quite self explaining:
An externally hosted image should be here but it was not working when we last tested it.
Eva, what do you think, can your schematic in the previous post be used with UCC2808 current mode IC?
UCC2808 datasheet
UCC2808 datasheet
If using op-amps in the feedback loop, will those op-amps need to be special, low-offset, low-drift devices?
( or some of those zero-offset, chopper-stabilized parts?)
Will an op-amp whose offset voltage drifts around, cause inaccuracies in regulation, or is DC input drift inherently cancelled by the feedback loop?
( or some of those zero-offset, chopper-stabilized parts?)
Will an op-amp whose offset voltage drifts around, cause inaccuracies in regulation, or is DC input drift inherently cancelled by the feedback loop?
IMHO typically not an issue, almost any sort of opamp has good enough offset specs compared to all other variables in circuit. IE your voltage reference might be 5% tolerance and so on, whereas some 5mV offset voltage is neglible to 5% tolerance of reference.funberry said:
Case is different if you require 0.01% output accuracy over time and temp, but such is rarely needed from power supplies.
Danko:
This arrengement is not directly useable with UCC2808 because its error amp has its non-inverting input internally tied to a 2V reference thus it only has the inverting input available.
Anyway, the dataseet shows the correct phototransistor arrangement, look how the H11A1 optocoupler is connected. It pulls up the inverting input instead of pulling down the non-inverting one, and those two 4.99K resistors ensure unity gain while the 432 ohm resistor ensures proper phototransistor biasing (it seems quite low a value, though).
The photodiode section is directly useable, though. My arrengement shows just how to use a discrete error-amp when a TL431 is not practical due to too high voltages (altough a TL431 would provide better performance).
This arrengement is not directly useable with UCC2808 because its error amp has its non-inverting input internally tied to a 2V reference thus it only has the inverting input available.
Anyway, the dataseet shows the correct phototransistor arrangement, look how the H11A1 optocoupler is connected. It pulls up the inverting input instead of pulling down the non-inverting one, and those two 4.99K resistors ensure unity gain while the 432 ohm resistor ensures proper phototransistor biasing (it seems quite low a value, though).
The photodiode section is directly useable, though. My arrengement shows just how to use a discrete error-amp when a TL431 is not practical due to too high voltages (altough a TL431 would provide better performance).
hi nazcalitonazcalito said:
in 300w at supply 60-0-60,you can volt peak +/-60vp or vrms=42.4v and you can power max at 8 ohm= 224 Wrms ( no lost in power amp)but class AB lost 35-40% or class d lost 10-20% and you have power in class AB at supply 60-0-60 V =134-145 w at 8 ohm or power in class D =175-201 w at 8 ohm
alab-pro.com
alabpro:
I'm looking at these SMPS:
alab pro SMPS
But I don't see any prices on this website.
Are these items for sale?
If so, where are the prices?
Does anyone think a potential customer could even remotely consider a product without knowing the price?
I'm looking at these SMPS:
alab pro SMPS
But I don't see any prices on this website.
Are these items for sale?
If so, where are the prices?
Does anyone think a potential customer could even remotely consider a product without knowing the price?
Eva,
Sorry to have taken so long to analyze this thing - and I am still at it - but, between work and my kids it is hard to find the time. OK, so much for digressing . . .
more feedback questions:
Your circuit compares the output (Vo) with a threshold (Vt) voltage. When the Vo > Vt, the circuit is on, and current flows through the Opto LED, and the current linearly increases with Vo. For Vo < Vt, the circuit is off and no current flows in the LED.
Assuming the pwm operates with maximum duty cycle for Verr is at its max - then when Vo < Vt the pwm runs at maximum duty cycle. When Vo > Vt, the duty cycle drops as Vo increases. Correct so far?
If so, then there is a design issue at that transition (at Vo = Vt). Wouldn't the desired operation see the pwm to go from max duty cycle to min duty cycle smoothly? So, the circuit needs to accomodate that.
any thoughts? I am on the right track?
Oh, one other thing. This circuit keeps the pwm running at max when Vo < Vt. That keeps the minimum voltage Vo, at Vt - hopefully. The top range of Vo is set by the transformer turns ratio. But say the design calls for Vo max to be less than the absolute max set by the turns ratio. Then, I believe, the feedback circuit needs to be designed such that the pwm goes to 0% duty cycle for Vo > Vmax desired. Seems correct to me - but am looking for the nod from you smps gurus
gene
Sorry to have taken so long to analyze this thing - and I am still at it - but, between work and my kids it is hard to find the time. OK, so much for digressing . . .
more feedback questions:
Your circuit compares the output (Vo) with a threshold (Vt) voltage. When the Vo > Vt, the circuit is on, and current flows through the Opto LED, and the current linearly increases with Vo. For Vo < Vt, the circuit is off and no current flows in the LED.
Assuming the pwm operates with maximum duty cycle for Verr is at its max - then when Vo < Vt the pwm runs at maximum duty cycle. When Vo > Vt, the duty cycle drops as Vo increases. Correct so far?
If so, then there is a design issue at that transition (at Vo = Vt). Wouldn't the desired operation see the pwm to go from max duty cycle to min duty cycle smoothly? So, the circuit needs to accomodate that.
any thoughts? I am on the right track?
Oh, one other thing. This circuit keeps the pwm running at max when Vo < Vt. That keeps the minimum voltage Vo, at Vt - hopefully. The top range of Vo is set by the transformer turns ratio. But say the design calls for Vo max to be less than the absolute max set by the turns ratio. Then, I believe, the feedback circuit needs to be designed such that the pwm goes to 0% duty cycle for Vo > Vmax desired. Seems correct to me - but am looking for the nod from you smps gurus
gene
EDITalabpro said:
HI EVA
-Output inductors = 6uH*3 and switching I test from Power amp Output power =64.8V AT 2OHM/CH AND TEST 33msON AND 66ms off full load
- The number of power devices =HGHT30N60C3D IGBT 63A600V FOR 4200W
- AND transformer size FOR 2400W=E55/28/21 OR ETD49
-FOR 3600W=E55/28/25 OR ETD54
-FOR 4200W =ETD54 AND UP 6500W=ETD59
-heatsink size yes very small but I use fan for cooling
and pls. check http://www.camcoaudio.com/pdf/tecton_p-partnertest.pdf he can made
-and PFC I use EE55 For PFC and ETD59 For swittching
Than you for response. I will check back in 2-3 days.
At first I thought "Why diagonal heatsink?"
Then I realize, that you can orient a fan on x axis or Y-axis and still get 45 degree airflow into the heat sink.
If heat sink aligned straight, then you are forced to orient fan only one way.
All heatsinks should be diagonal!
Well, a class B or AB amplifier with +-64.8V rails driving a 2 ohm resistive load produces no more than 1000Wrms before clipping. Two channels would add up to 2000Wrms. Assuming 70% effiency, total average power consumption would be approx 2850W.
Finally, these 2850W drawn during 33ms of each 99ms interval yield a total average power consumption of 950W, and this is the actual continuous power rating of that power supply, despite the fact that it could provide 4000W transients. This is more than enough for audio signals, though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.