Re: Post #1983 and Post #2012
Hari,
Something just struck me about the PDF you posted in Post #2012, then I checked Post #1983. Both of the PDFs in Post #2012 and Post #1983 appear to be smaller, perhaps about 10-20% the size, of the PDF I eMailed you after I was able to reduce the PDF to about 64Kb. If you can print both the PDF I eMailed you, one of the PDFs you created and then print from Protel your PCB image to confirm all the images print the correct PCB image size. You have added C11B and C13B since the PDF I eMailed you so the PCB will be a bit longer, but I think the PCB width should be the same. I have compared the width of the TO-247's and they seem smaller in the PDFs you have created than the one I created. If you have a TO-247 device handy you can also confirm which, if any, of the PDFs print the right size footprint for the TO-247 and match the leads pads for the TO-247 on the printouts.
Regards,
John L. Males
Willowdale, Ontario
Canada
15 July 2007 19:44
Official Quasi Thread Researcher
jethari said:
Hari,
Something just struck me about the PDF you posted in Post #2012, then I checked Post #1983. Both of the PDFs in Post #2012 and Post #1983 appear to be smaller, perhaps about 10-20% the size, of the PDF I eMailed you after I was able to reduce the PDF to about 64Kb. If you can print both the PDF I eMailed you, one of the PDFs you created and then print from Protel your PCB image to confirm all the images print the correct PCB image size. You have added C11B and C13B since the PDF I eMailed you so the PCB will be a bit longer, but I think the PCB width should be the same. I have compared the width of the TO-247's and they seem smaller in the PDFs you have created than the one I created. If you have a TO-247 device handy you can also confirm which, if any, of the PDFs print the right size footprint for the TO-247 and match the leads pads for the TO-247 on the printouts.
Regards,
John L. Males
Willowdale, Ontario
Canada
15 July 2007 19:44
Official Quasi Thread Researcher
Re: Re: Post #1983 and Post #2012
Hari,
I have done some more research sourounding your PCB image size and the C11/C13 capacitors that have been concerning me for a while now.
I have compared your two posted PCBs to the NMOS200 TO-247 version quasi posted near the start of the year. I find both of your posted PCBs to have the TO-247 footprint smaller than the NMOS200 TO-247 and the PCB I emailed you. The NMOS200 TO-247 and the PCB I eMailed you TO-247 footprints are the same size.
Regarding the C11/C13 rail filter capacitors, I have had this feeling their footprint was too small due to the small dimension of 63V. I checked the quasi "Classic" PCB and the footprints for your C11 and C13 are smaller. C11 and C13 are about 23-25mm in diameter on the quasi "Classic" PCB. I measured the diameter of C11 and C13 on the image I used to create the PDF I eMailed you and C11 and C13 are 16mm in diameter. It may be for the compact PCB nature to use 63V capacitors for the single pair NMOS TO-247, but I think with two pairs of TO-247's one can certainly run rails at 63V or tad above with two pairs of TO-247s. You did indicate you wanted to create a "L" mounting PCB like the NMOS200, but be like the quasi "Classic" in all other respects. Perhaps you might want to consider using 200V 330uF's for C11/13/A/B. I have some APT4025BNs that derated can handle 200W each derated to 60C and I know of other APT devices and some IRF devices like IRFP250/460/260 that would need a higher rated C11/C13 working voltage. One of the reasons I created the single pair and two pair versions of the "classic" PCB was due to these higher power TO-247 devices quasi said would be interesting to use on the NMOS200 TO-247 PCB. Using two pairs of APT4025BNs allows for 400W which would be 200W/8 Ohms and 400W/4 Ohms derated to 60C. It seems that quasi has dimensioned these filter capacitors at greater than 2 x Vrail on the "Classic" NMOS. I am not sure if this was a design decision or based on what was in the quasi store in light of the NMOS200 C11/C13 being rated at 63V compared to 200V on the "Classic" PCB. I do know that the output MOSFET devices need a rating of 2 x VRail plus some spare per the answer I had in Post #1920. I suspect C11/C13 ideally would need to be the same for the same reasons the output MOSFETs need to have a Vdss rating of greater than 2 x VRail. That said, for the quasi NMOS200 designed for rails up to 50V (and Post #1337) and C11/C13 dimensioned at 63V.
Regards,
John L. Males
Willowdale, Ontario
Canada
15 July 2007 (19:45 -) 22:03
Official Quasi Thread Researcher
keypunch said:
Hari,
I have done some more research sourounding your PCB image size and the C11/C13 capacitors that have been concerning me for a while now.
I have compared your two posted PCBs to the NMOS200 TO-247 version quasi posted near the start of the year. I find both of your posted PCBs to have the TO-247 footprint smaller than the NMOS200 TO-247 and the PCB I emailed you. The NMOS200 TO-247 and the PCB I eMailed you TO-247 footprints are the same size.
Regarding the C11/C13 rail filter capacitors, I have had this feeling their footprint was too small due to the small dimension of 63V. I checked the quasi "Classic" PCB and the footprints for your C11 and C13 are smaller. C11 and C13 are about 23-25mm in diameter on the quasi "Classic" PCB. I measured the diameter of C11 and C13 on the image I used to create the PDF I eMailed you and C11 and C13 are 16mm in diameter. It may be for the compact PCB nature to use 63V capacitors for the single pair NMOS TO-247, but I think with two pairs of TO-247's one can certainly run rails at 63V or tad above with two pairs of TO-247s. You did indicate you wanted to create a "L" mounting PCB like the NMOS200, but be like the quasi "Classic" in all other respects. Perhaps you might want to consider using 200V 330uF's for C11/13/A/B. I have some APT4025BNs that derated can handle 200W each derated to 60C and I know of other APT devices and some IRF devices like IRFP250/460/260 that would need a higher rated C11/C13 working voltage. One of the reasons I created the single pair and two pair versions of the "classic" PCB was due to these higher power TO-247 devices quasi said would be interesting to use on the NMOS200 TO-247 PCB. Using two pairs of APT4025BNs allows for 400W which would be 200W/8 Ohms and 400W/4 Ohms derated to 60C. It seems that quasi has dimensioned these filter capacitors at greater than 2 x Vrail on the "Classic" NMOS. I am not sure if this was a design decision or based on what was in the quasi store in light of the NMOS200 C11/C13 being rated at 63V compared to 200V on the "Classic" PCB. I do know that the output MOSFET devices need a rating of 2 x VRail plus some spare per the answer I had in Post #1920. I suspect C11/C13 ideally would need to be the same for the same reasons the output MOSFETs need to have a Vdss rating of greater than 2 x VRail. That said, for the quasi NMOS200 designed for rails up to 50V (and Post #1337) and C11/C13 dimensioned at 63V.
Regards,
John L. Males
Willowdale, Ontario
Canada
15 July 2007 (19:45 -) 22:03
Official Quasi Thread Researcher
Originally posted by Keypunch
Regarding the C11/C13 rail filter capacitors, I have had this feeling their footprint was too small due to the small dimension of 63V. I checked the quasi "Classic" PCB and the footprints for your C11 and C13 are smaller.
John
You're right there. I did not realize that for the nmos350 I would need higher voltage caps. I have now changed to a single 680/100V cap of 16mm dia. I checked some datasheets for caps and this dia is ok for this value.
Thanks
Hari
Regarding the C11/C13 rail filter capacitors, I have had this feeling their footprint was too small due to the small dimension of 63V. I checked the quasi "Classic" PCB and the footprints for your C11 and C13 are smaller.
John
You're right there. I did not realize that for the nmos350 I would need higher voltage caps. I have now changed to a single 680/100V cap of 16mm dia. I checked some datasheets for caps and this dia is ok for this value.
Thanks
Hari
Hi,
it's not really the NMos350 that needs the higher voltage caps.
It's the transformer voltage that determines the cap voltage.
40Vac is the limit for 63Vdc caps.
50Vac is the limit for 80Vdc caps.
63Vac is the limit for 100Vdc caps.
Some builders use a higher margin than I do, so decide if you want to set your limits lower than suggested above.
it's not really the NMos350 that needs the higher voltage caps.
It's the transformer voltage that determines the cap voltage.
40Vac is the limit for 63Vdc caps.
50Vac is the limit for 80Vdc caps.
63Vac is the limit for 100Vdc caps.
Some builders use a higher margin than I do, so decide if you want to set your limits lower than suggested above.
AndrewT said:
Andrew
What I meant was that as I was originally modifying the nmos200 pcb layout to fit the nmos 350 schematic, I overlooked the cap voltage ratings. I will be using 65V DC rails and hence the choice of 100V caps. The limits you mention is an excellent guideline.
Thanks
Hari
Hi jethari,
Please disregard this if it has been mentioned, explained or corrected before, just took a quick look.
Your decoupling caps C9, C10, C11 and C13 they all share an incredibly tiny ground trace while you have lots of PCB space available, why?
The loss in that trace will make decoupling less effective due to added trace resistance and C5 and C6 connected halfway will have bad ground reference.
jethari said:
Please disregard this if it has been mentioned, explained or corrected before, just took a quick look.
Your decoupling caps C9, C10, C11 and C13 they all share an incredibly tiny ground trace while you have lots of PCB space available, why?
The loss in that trace will make decoupling less effective due to added trace resistance and C5 and C6 connected halfway will have bad ground reference.
4fun
This point is certainly NOT to be disregarded. Never thought of it that way. As you say there is definitely enough space for a larger ground trace. The only weak link would be the two wire links (no pun intended!). I guess that can be overcome by using pretty thick wire for the links.
Thanks
Hari
This point is certainly NOT to be disregarded. Never thought of it that way. As you say there is definitely enough space for a larger ground trace. The only weak link would be the two wire links (no pun intended!). I guess that can be overcome by using pretty thick wire for the links.
Thanks
Hari
Quasi
I came across this mosfet from ST called STW20NM60. This is a 20A 600V device and the description on the datasheet is identical to that of the IRFP450. Would this be a viable alternative to the IRFP450? Maybe even a single pair version using 60V DC rails to drive 8 ohm speakers?
Thanks
Hari
I came across this mosfet from ST called STW20NM60. This is a 20A 600V device and the description on the datasheet is identical to that of the IRFP450. Would this be a viable alternative to the IRFP450? Maybe even a single pair version using 60V DC rails to drive 8 ohm speakers?
Thanks
Hari
Hari,
The + trace of C11 to the + Rail needs to be much wider. Given the narrowist part of the + rail trace is about 1/2 of where C11 + trace goes to the + rail trace I would suggest there is lots of opportunity to make the C11 + trace as wide as the narrowist part of the + rail trace. That said then the C13 - trace to the -Rail should be made the same width as the C11 + trace will made.
Also, the narrowist part of the + and - rail traces should be the same width.
I think L1 was better off not being moved. Now you may have stray L1 related currents floating on one side of the output devices unless you wire link back to R35 point.
Regards,
John L. Males
Willowdale, Ontario
Canada
17 July 2007 22:10
Official Quasi Thread Researcher
The + trace of C11 to the + Rail needs to be much wider. Given the narrowist part of the + rail trace is about 1/2 of where C11 + trace goes to the + rail trace I would suggest there is lots of opportunity to make the C11 + trace as wide as the narrowist part of the + rail trace. That said then the C13 - trace to the -Rail should be made the same width as the C11 + trace will made.
Also, the narrowist part of the + and - rail traces should be the same width.
I think L1 was better off not being moved. Now you may have stray L1 related currents floating on one side of the output devices unless you wire link back to R35 point.
Regards,
John L. Males
Willowdale, Ontario
Canada
17 July 2007 22:10
Official Quasi Thread Researcher