Hi,
Have you ever tried using SMD resistors for MOSFET source resistors?
Chip Resistor - Surface Mount | Resistors | DigiKey
They are rated 3W, for 0.1R source resistors the power rating is sufficient. The power rail fuse rating is only 7A so if 7A goes through 0.1R it will dissipate 4.9W. And that is the maximum power. For normal music playing, the power will be a lot less.
My worry is that I don't know if the power rating can only be realised by soldering the resistors on large copper areas to dissipate heat. If only one side of the resistor is soldered on some small area of copper (track of 5.5mm wide and 10mm long), will the power rating be much reduced and the resistors get burnt?
Have you ever tried using SMD resistors for MOSFET source resistors?
Chip Resistor - Surface Mount | Resistors | DigiKey
They are rated 3W, for 0.1R source resistors the power rating is sufficient. The power rail fuse rating is only 7A so if 7A goes through 0.1R it will dissipate 4.9W. And that is the maximum power. For normal music playing, the power will be a lot less.
My worry is that I don't know if the power rating can only be realised by soldering the resistors on large copper areas to dissipate heat. If only one side of the resistor is soldered on some small area of copper (track of 5.5mm wide and 10mm long), will the power rating be much reduced and the resistors get burnt?
Last edited:
From the data sheet, it appears they want both leads treated the same way:
"The WSR3 requires a minimum of 1050 sq. mil. circuit traces connecting to the recommended solder pad."
http://www.vishay.com/docs/30101/wsr.pdf
Thanks a million. I was trying to spot that line but I failed!
1050 sq. mil? Is mil millimetre or thousandth of an inch?
It doesn't make sense if it is thousandth of an inch because it is only 0.68 sq millimetres which is really too small.
If it is millimetre then it requires a 33mm x 33mm pad which seems to be quite large.
1050 sq. mil? Is mil millimetre or thousandth of an inch?
It doesn't make sense if it is thousandth of an inch because it is only 0.68 sq millimetres which is really too small.
If it is millimetre then it requires a 33mm x 33mm pad which seems to be quite large.
Last edited:
Think I'll check with Vishay about this, these parts look useful.
Last edited:
I´m never comfortable placing a power resistor flat against a PCB, always leave at least a couple mm (and preferrably 5mm or more) between them ... which is impossible for a SMD part, by definition.
Reading the datasheet, their normal unit is Imperial, with Metric indicated between parentheses, so mil must be thousandths of an inch ... by the way the old standard for PCB dimensions.
Each pad is quite large 4mm by 6mm and there´s two of them, the 1050 sq mil referred to tracks does not make much sense, since they travel far away.
I think tracks must not contribute much to dissipation more than , say, 10 or 20 mm away so their total surface should be irrelevant.
If anything they might suggest minimum track width leading to/from pad, worst case also suggest copper thickness.
Reading the datasheet, their normal unit is Imperial, with Metric indicated between parentheses, so mil must be thousandths of an inch ... by the way the old standard for PCB dimensions.
Each pad is quite large 4mm by 6mm and there´s two of them, the 1050 sq mil referred to tracks does not make much sense, since they travel far away.
I think tracks must not contribute much to dissipation more than , say, 10 or 20 mm away so their total surface should be irrelevant.
If anything they might suggest minimum track width leading to/from pad, worst case also suggest copper thickness.
The issue is inductance. Through hole power resistors are usually wirewound type and they can measure fairly high inductance. SMD resistors are usually non-inductive.
Given that the MOSFET source resistors would probably dissipate less than 1W when playing music I think I could give the SMD a go.
The Zobel power resistor should also be of non-inductive type. I used a SMD resistor for the Zebel soldered on fairly small copper tracks and it survived without problems when playing music, until one day I measured the amplifier with 10kHz square wave with a 8R dummy load and the SMD resistor smoked. I replaced with a wirewound and it seemed to work OK.
Given that the MOSFET source resistors would probably dissipate less than 1W when playing music I think I could give the SMD a go.
The Zobel power resistor should also be of non-inductive type. I used a SMD resistor for the Zebel soldered on fairly small copper tracks and it survived without problems when playing music, until one day I measured the amplifier with 10kHz square wave with a 8R dummy load and the SMD resistor smoked. I replaced with a wirewound and it seemed to work OK.
This is a 3W 6.8R metal Element through-hole I could use for the Zobel. This one is fairly cheap too. But when I thought of it what is the difference between it and a SMD? It uses the same material. The only advantage with the through-hole is that it could be lifted 2mm from the PCB so that it has more air surface for heat dissipation. But if I use larger pads perhaps the SMD could be as good. or not?
ROX3SJ6R8 - TE CONNECTIVITY / NEOHM - Through Hole Resistor, 6.8 ohm, 350 V, Axial Leaded, 3 W, ± 5%, ROX Series | element14 Australia
ROX3SJ6R8 - TE CONNECTIVITY / NEOHM - Through Hole Resistor, 6.8 ohm, 350 V, Axial Leaded, 3 W, ± 5%, ROX Series | element14 Australia
Regularly do SMD boards with high power devices placed flat to the board, but these are multi layer designs where I do a lot of thermal engineering as well with the design. Thermal via's, multiple planes, 2oz copper all layers. The important thing is having either thermal pads to heat-sinks or some method of spreading the heat.
The size of the track to the device is going to be at least 0.200 wide for current capacity and good design practice so the 1050 sq mil trace is easily achievable, even a 0.100 wide track would do the trick....
I sent an enquiry to Vishay Dale Resistors Division and they sent me a reply.
They said: "You are correct that the square mil = 0.001". They also attached a Power vs Temperature curve but I cannot attach it here with my work computer.
It is good that manufacturer gives a reply.
So I think I can use the SMD resistor even with small tracks to the pads.
They said: "You are correct that the square mil = 0.001". They also attached a Power vs Temperature curve but I cannot attach it here with my work computer.
It is good that manufacturer gives a reply.
So I think I can use the SMD resistor even with small tracks to the pads.
I use the 3W Dale SMDs regularly in power supplies by soldering on heavy guage copper 'legs' for thru hole and open wiring use - they are readily available as 0.1 and so forth, so can add them in series - common value source resistor 0.47 = 3 x 0.1 + 0.15R, etc - I'm not so keen on the sound of the Metal Oxides for power resistors but this maybe just my prejudice (of which I have too many!)
As usual, that magic term 'it depends' comes to the fore - if you're thinking about using them for big power amps, then possibly they won't have enough heat capacity even if you make up the total using 0.1R devices in series - on the other hand, for low power classA amps, the source current is down in the 1 - 2Amp area and thus the power dissipation much lower
On some amps that run a lot of negative feedback, it's probably just as good to use generic metal oxide source resistors - Greg's GB150D works okay with these and no discernible difference when replaced/upgraded with Koa metal films - the 'Blameless' had a similar characteristic, IMO anyway - I'm no guru.
However, on the F5 amp, for example, it's a very simple circuit so any different components, including the power supply, has immediate effects on the sound - I use the traditional tester, the AKG k701 headphones, to avoid confusion as it's easy to fool yourself with some combinations with initial impressions and all that ...
Hope this is of some use ....
On some amps that run a lot of negative feedback, it's probably just as good to use generic metal oxide source resistors - Greg's GB150D works okay with these and no discernible difference when replaced/upgraded with Koa metal films - the 'Blameless' had a similar characteristic, IMO anyway - I'm no guru.
However, on the F5 amp, for example, it's a very simple circuit so any different components, including the power supply, has immediate effects on the sound - I use the traditional tester, the AKG k701 headphones, to avoid confusion as it's easy to fool yourself with some combinations with initial impressions and all that ...
Hope this is of some use ....
I guess I can appreciate what you found as I previously found that changing components, like, caps, or any circuitry components could make a huge difference in sound. Now I think they could possibly be due to noise emission of one kind or another. If all current loops / returns are kept at a minimum using planes / twisted pairs, etc, the effects of components seem to be reduced by a lot. I have not been able to tell the difference between resistors (never used carbon resistors), perhaps I have not really tried them.
It's an interesting thing but I've found that some people can't hear differences in system components at all, be it interconnects, caps in line, caps in power supplies, resistors, passive crossover upgrades (won't go there!) and so on - they just seem to listen for other different things and ignore all the 'kerfuffle' about components - people do have rather different requirements in musical reproduction as distinct from performing artists
As you would expect, it's the capacitors that produce the bigger sound differences but with the resistors it seems to be a combination effect - the difference of just one 'fancy' resistor in the signal path is small but changing a number of them or getting rid of poor quality ones seems to generate more significant results particularly in feedback networks, high gain stages, etc
A few years ago, I went thru a period of experimenting with various types of resistor combinations on simple preamp and power amps and, as you mentioned above, altering the quality of the pcbs, better wiring, etc has just as much contribution to improved sonics as the components if not more - IMO, here mind you.
Years ago, the Roedestein 'resista' version of the metal film resistors were regarded as 'about the best' for a European resistor - this is about 30 years ago mind you, before the internet - today we have all sorts of types easily available - I went thru a period of changing components using those 'naked vishays' (TC2575s), saw the light (and the empty chequebook) and resorted to Susumis and now I generally start with Koa metal films if I have them and occasionally use the Rhopoints (lower values made up here from Manganin wire) - it's a bit of an interest 'thing' but surprising how much a difference it can make
We have the Melb Audio Show here next w/end at the Rialto - lots of silly $ items about....
As you would expect, it's the capacitors that produce the bigger sound differences but with the resistors it seems to be a combination effect - the difference of just one 'fancy' resistor in the signal path is small but changing a number of them or getting rid of poor quality ones seems to generate more significant results particularly in feedback networks, high gain stages, etc
A few years ago, I went thru a period of experimenting with various types of resistor combinations on simple preamp and power amps and, as you mentioned above, altering the quality of the pcbs, better wiring, etc has just as much contribution to improved sonics as the components if not more - IMO, here mind you.
Years ago, the Roedestein 'resista' version of the metal film resistors were regarded as 'about the best' for a European resistor - this is about 30 years ago mind you, before the internet - today we have all sorts of types easily available - I went thru a period of changing components using those 'naked vishays' (TC2575s), saw the light (and the empty chequebook) and resorted to Susumis and now I generally start with Koa metal films if I have them and occasionally use the Rhopoints (lower values made up here from Manganin wire) - it's a bit of an interest 'thing' but surprising how much a difference it can make
We have the Melb Audio Show here next w/end at the Rialto - lots of silly $ items about....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.