Hypex Linear PSU DC-error protection

[doctordata]

Yes, I can steal it from 2x13Vac buffer PSU trafo and requlate to lower voltage.

Hi Sami ,

Did some work today ,

Downloaded Eagle to draw some schematics and LTspice , need to get familiar with it again, cause I didn't use it for a long time .
Found some nice FET's IRFB4115PBF 150V 11 mOhm but have to check all the specs . The Picaxe chip needs 5 Volts low current , so a 78l05 regulator will do .
The PVI you mentioned is SMD , not a problem ?
Had some thoughts about the software for the PIC .
Would check for primary FET voltage first on startup , then check voltage after the FET's . If there is any , Fault condition, FET's are shorted . Blink LED
No voltage , open FET's ,check voltage , if OK go to DC fault loop , check for low signal on DC protect output on UCD , if so close FET's blink LED .
If you have any other thoughts let me know .

Cheers ,

Rens

 

[Piisami]

The PVI you mentioned is SMD , not a problem ?

SMD is not a problem

Would check for primary FET voltage first on startup , then check voltage after the FET's . If there is any , Fault condition, FET's are shorted . Blink LED
No voltage , open FET's ,check voltage , if OK go to DC fault loop , check for low signal on DC protect output on UCD , if so close FET's blink LED .
If you have any other thoughts let me know

This sounds just right

I'm going to have an soft-start for big trafos and capacitors, so there could be a startup delay for couple of seconds to open the FETs.

 

[doctordata]

SMD is not a problem



This sounds just right

I'm going to have an soft-start for big trafos and capacitors, so there could be a startup delay for couple of seconds to open the FETs.

The soft start and delay can be handled by the picaxe as well very flexible and programmable . Would be good , to control the soft start by the Pic , so it could also disconnect the start relay when a FET is diagnosed to be shorted .
The AUX power supply for the pic should be always on for most program options ( after the power switch )
Can you order parts from Digikey in your country ?

Cheers ,

Rens

 

[Piisami]

The soft start and delay can be handled by the picaxe as well very flexible and programmable . Would be good , to control the soft start by the Pic , so it could also disconnect the start relay when a FET is diagnosed to be shorted .

Cool..

Can you order parts from Digikey in your country ?

Yes, I can.

 

[doctordata]

This schematic should do the trick .
Could not find a PV symbol in Eagle ( Which one did you use ?), but think this is clear enough .
Did not have enough pins to serve the soft start , but think there are plenty available out there .
The picaxe senses the DC error signal and the positive and negative rail for FET failure .
Will calculate the values the coming days for the different UCD's and write some code as we discussed before and test the circuit .

Cheers ,

Rens

 

[doctordata]

schematic revsion

Small error in schematic ,
R8 should be connected to ground , not +5 V

Cheers ,

Rens

 

[twest820]

I was thinking to parallel two or more FETs, this would make more choice for FETs and, to my knowledge, lower the FET resistance.

Consider IRF's e-fuse lineup or similar parts; at ~1mOhm Rds_on the supply impedance increase from inserting protection circuitry is going to be governed by the changes needed to hook up the extra FETs.

This thread's been about DC protection so far, but as doctordata is touching on, including overcurrent protection is something to think about. I'm not intimately familiar with Hypex's product line but Hypex doesn't call out overcurrent protection in the 400HG module datasheet and they neglect to specify a trip limit for 180HG and 700HG. So, depending on application, a lower master limit on the supply or per channel protection may desirable.

Main limitation with low Rds_on FETs for protection is protection features on hot swap controllers typically trip at 50mV. So if you want fairly turnkey circuit breaker type overcurrent protection using ICs with Rds_on sensing and an integrated charge pump for an high side driver a higher sense resistance is needed. Adding ~10mOhm sense resistor + ~2mOhm FET to the supply impedance isn't really a big deal with a linear supply, though---for example, a typical 10,000uF electrolytic is around 180 mOhm +- 32 mOhm at 100Hz---and tripping off a sense resistor is generally more accurate. If you don't mind the programming VOM1271s and a micro aren't a bad way to go, though; once tripped it'd take 1.25ms to turn off an IRFH8201.

I think the output impedance is mainly set by the DC resistance of the output inductor and the Rds on of the output FETS, but I'm getting on thin ice here

Depends on parts selection and amp topology. Rds_ons of audio optimized FETs tend be an order of magnitude higher than the ESR of class D specific inductors. In post-filter implementations the output impedance of the supply, switch, and filter is divided by the loop gain (about 30dB for UcD). In pre-filter implementations output impedance goes to "infinity" at filter resonance, meaning increasing output impedance with audio frequency which depends on the Q and resonant frequency of the parts selected.

 

[doctordata]

Consider IRF's e-fuse lineup or similar parts; at ~1mOhm Rds_on the supply impedance increase from inserting protection circuitry is going to be governed by the changes needed to hook up the extra FETs.

Nice parts , thanks for mentioning but not suitable for this application as the max voltage of this device is 25V . Would be nice for low voltage applications though !

For the over current protection , the ACS712 current sensors can be implemented in the schematic instead of the resistor network on the output voltage , in series with the FET's . They have a Max resistance of 1.5 mOhms!
and produce 66mV per ampere , easy measurable with the picaxe .
Only problem could be the picaxe's speed ( due to it's build in Basic interpreter ), running at 32Mhz it will take about halve a millisecond to detect over current . Fast enough ?

Can post schematic if anybody is interested .

Cheers ,

Rens

 

[twest820]

Hence the or similar parts bit.

I happen to prefer 3.3V micros so would use the FAULT outputs from pair of ACS709s as interrupts if I didn't implement a hot swap controller based board. That turns the micro into a DC level integrator followed by a three input latching NOR, which is easy enough to implement with a few discretes and a little combinatorial 7400 logic. Otherwise, eh, might as well make it a protection cape/shield/add on board/whatever---in DIY quantity it costs as much to put a micro on a board as it does to buy an LPC-Link2 or the like.

The ACS762 and the earlier ACS760 are rather interesting as well but don't seem to be (or have been) available in small enough quantities for distributors to stock.

 

[doctordata]

Hence the or similar parts bit.

I happen to prefer 3.3V micros so would use the FAULT outputs from pair of ACS709s as interrupts if I didn't implement a hot swap controller based board. That turns the micro into a DC level integrator followed by a three input latching NOR, which is easy enough to implement with a few discretes and a little combinatorial 7400 logic. Otherwise, eh, might as well make it a protection cape/shield/add on board/whatever---in DIY quantity it costs as much to put a micro on a board as it does to buy an LPC-Link2 or the like.

The ACS762 and the earlier ACS760 are rather interesting as well but don't seem to be (or have been) available in small enough quantities for distributors to stock.

Did you ever read the word simple in the previous posts ?
Hot swap controller in an Audio Amp ?

Cheers ,

Rens

 

[twest820]

You're saying two Hall effect sensors, an ADC and its input buffer, a pair of optoisolated gate drivers, and firmware is simpler than a turnkey IC which does the same thing?

 

[doctordata]

FAst turn off at DC and overcurrent protection

You're saying two Hall effect sensors, an ADC and its input buffer, a pair of optoisolated gate drivers, and firmware is simpler than a turnkey IC which does the same thing?

ADC and buffer are all in the PIC . Firmware is just a couple of lines basic code .
The attached schematic shows the ACS712, available in 5 , 20 and 30Amps version .

Cheers ,

Rens

 

[doctordata]

VOM1271 pinout

Forgot to blank the pin numbers of the PV .
They are not correct , just used some other part for the schematic .
For correct pinout , consult the datasheet

 

[Piisami]

I've done some work on the past few days.

Here is the full schematic, including the main PSU, driver PSU and buffer psu:
https://www.dropbox.com/meta_dl/eyJzdWJfcGF0aCI6ICIiLCAidGVzdF9saW5rIjogZmFsc2UsICJzZXJ2ZXIiOiAiZGwuZHJvcGJveHVzZXJjb250ZW50LmNvbSIsICJpdGVtX2lkIjogbnVsbCwgImlzX2RpciI6IGZhbHNlLCAidGtleSI6ICIxc2VwMzR3MjRvenVwM2sifQ/AAI54Gm2_Ii12zzL9GkwzFCWymSF1dHB3MSrh6N4ryNwZA?dl=1

And here is the board, excluding the buffer PSU, which is going to be on a separate board:
https://www.dropbox.com/meta_dl/eyJzdWJfcGF0aCI6ICIiLCAidGVzdF9saW5rIjogZmFsc2UsICJzZXJ2ZXIiOiAiZGwuZHJvcGJveHVzZXJjb250ZW50LmNvbSIsICJpdGVtX2lkIjogbnVsbCwgImlzX2RpciI6IGZhbHNlLCAidGtleSI6ICI2ODI1YThwMndmYTNvZTIifQ/AAIhNPRTFCtO8PzOlpxeUekzAObA9hDUPnrCi9kvIwBAiQ?dl=1

And here are the relevant files:
https://www.dropbox.com/meta_dl/eyJzdWJfcGF0aCI6ICIiLCAidGVzdF9saW5rIjogZmFsc2UsICJzZXJ2ZXIiOiAiZGwuZHJvcGJveHVzZXJjb250ZW50LmNvbSIsICJpdGVtX2lkIjogbnVsbCwgImlzX2RpciI6IGZhbHNlLCAidGtleSI6ICJ0anU3OXMxMXZlMGQ5cnIifQ/AAJkqN7HUgPJw8rmX8RpOqxpyuYeVS1LEZLnehqhw_UMjA?dl=1

I have a couple of concerns / questions:
-I will take the dc voltage for the PIC regulator from the buffer PSU positive rail, is that OK?
-The auto on (/ON) circuitry should be ok, but the is the AC connected correctly for the diodes D12+D13?
-I'm really suspicious about ACS712. A small SMD component should handle that much power I would like the power rail traces to be as wide as possible, but it is not possible with those tiny components...
-The fuses may be omitted? there will be a fuse on the main socket anyway..
-Also, you may suggest proper rectifier diodes

Any other comments and suggestions are welcome

BR, Sami

 

[Piisami]

Also I went crazy and paralleled the FETs because I want to lower the resistance and to have plenty of headroom.

I will also add more capacitors with lower capacitance to lower the ESR.

 

[doctordata]

I've done some work on the past few days.

Here is the full schematic, including the main PSU, driver PSU and buffer psu:
https://www.dropbox.com/meta_dl/eyJ...m2_Ii12zzL9GkwzFCWymSF1dHB3MSrh6N4ryNwZA?dl=1

And here is the board, excluding the buffer PSU, which is going to be on a separate board:
https://www.dropbox.com/meta_dl/eyJ...RTFCtO8PzOlpxeUekzAObA9hDUPnrCi9kvIwBAiQ?dl=1

And here are the relevant files:
https://www.dropbox.com/meta_dl/eyJ...7HUgPJw8rmX8RpOqxpyuYeVS1LEZLnehqhw_UMjA?dl=1

I have a couple of concerns / questions:
-I will take the dc voltage for the PIC regulator from the buffer PSU positive rail, is that OK?
-The auto on (/ON) circuitry should be ok, but the is the AC connected correctly for the diodes D12+D13?
-I'm really suspicious about ACS712. A small SMD component should handle that much power I would like the power rail traces to be as wide as possible, but it is not possible with those tiny components...
-The fuses may be omitted? there will be a fuse on the main socket anyway..
-Also, you may suggest proper rectifier diodes

Any other comments and suggestions are welcome

BR, Sami

Had a quick look at the schematics , looks nice !
You connected R4 the wrong way Look at my schematics again .
It's a pull up resistor for the DC error signal .
I can't send you programmed SMD picaxe chips , have only DIP8
Not sure if the PV can drive 4 Parallel FET's , that's a lot of gate charge !
Have some more comments , but no time tonight .
Come back tomorrow .

Cheers ,

Rens

Cheers ,

 

[doctordata]

Had a quick look at the schematics , looks nice !
You connected R4 the wrong way Look at my schematics again .
It's a pull up resistor for the DC error signal .
I can't send you programmed SMD picaxe chips , have only DIP8
Not sure if the PV can drive 4 Parallel FET's , that's a lot of gate charge !
Have some more comments , but no time tonight .
Come back tomorrow .

Cheers ,

Rens

Cheers ,

Another quick look , IC4 Vcc to +5 Volt

 

[twest820]

I'm really suspicious about ACS712. A small SMD component should handle that much power I would like the power rail traces to be as wide as possible, but it is not possible with those tiny components.

I'm not following. How is dissipating less than a watt a problem? Since audio is relatively low voltage it should be easy to get below the 23 C/W of Allegro's recommended layout as isolation and creepage requirements are reduced and the pours can be wider. But even with a 50 C/W layout and 40C chassis temperature the junctions would be 65C at 20A RMS.

Why aren't you pouring both sides of the board and rotating and moving the ACS712s to get more copper? It's not like it matters which side of the FETs they're on either.

(I'd also suggest taking a look at CRC's lack of line regulation to see if that's really what you want.)

 

[doctordata]

over current protection

I've done some work on the past few days.

Here is the full schematic, including the main PSU, driver PSU and buffer psu:
https://www.dropbox.com/meta_dl/eyJ...m2_Ii12zzL9GkwzFCWymSF1dHB3MSrh6N4ryNwZA?dl=1

And here is the board, excluding the buffer PSU, which is going to be on a separate board:
https://www.dropbox.com/meta_dl/eyJ...RTFCtO8PzOlpxeUekzAObA9hDUPnrCi9kvIwBAiQ?dl=1

And here are the relevant files:
https://www.dropbox.com/meta_dl/eyJ...7HUgPJw8rmX8RpOqxpyuYeVS1LEZLnehqhw_UMjA?dl=1

I have a couple of concerns / questions:
-I will take the dc voltage for the PIC regulator from the buffer PSU positive rail, is that OK?
-The auto on (/ON) circuitry should be ok, but the is the AC connected correctly for the diodes D12+D13?
-I'm really suspicious about ACS712. A small SMD component should handle that much power I would like the power rail traces to be as wide as possible, but it is not possible with those tiny components...
-The fuses may be omitted? there will be a fuse on the main socket anyway..
-Also, you may suggest proper rectifier diodes

Any other comments and suggestions are welcome

BR, Sami

I read the datasheets of the UCD400OEM and 700OEM today and they DO have build in overcurrent protection ! So you can ditch the ACS712 and it's components and go back to my first schematic with the resister dividers on the output rails . ( I did enclose a picture how to implement high current traces with the ACS712 because it's a very nice chip !)
DC voltage for pic is OK
Your negative rectifiers D8-D11 and D22-D25 have the wrong polarity .
Mur860 is only 8 Amps better take MUR2020 or MUR2040 20 Amps 200 or 400V for the 55 AC lines .
You might consider using two PV's in parallel per side to be sure the FET's are switched on fully and fast enough . I'm not an expert in Fet formulas , maybe one of the smart guys here can do the math on driving four IRFB4115's with a single VOM1271

Cheers ,

Rens

 

[doctordata]

It's not like it matters which side of the FETs they're on either.

It matters , you want them on the secondary FET side to check if there is current flowing when the FET's are closed on startup and so detecting shorted FET's with the PIC.

Cheers ,

Rens