and how do you propose to machine/grind/sand the spotface to make it flat enough to conduct heat from the device, with minimal Rth s-a?
You can make your own end sander bit
Ofcourse you need a minimum of tools
But point is, theres no need for advanced expencive tools
Try to hold something straight against the device
Look up against a window or other light source, and you will see how plane that thing is, or rather isnt
how many channels are being powered by this transformer?
What is the transformer?
What is the continuous current draw from the PSU, i.e. output bias plus amplifier quiescent current?
I use a Talema toroidal for two channels, its getting hot when bias is 0.71V (when it had the insulators problems).
How can I check the psu continuous current and the amplifier quiescent current?
# Power rating: 300VA, 10A
# Core loss, typ: 1,7W
# Copper loss, typ (120 degrees celsius): 22,5W
# Size: 115x58mm
# Primary: 1x230V
# Secondary: 2x18V
Last edited:
300VA, 18+18Vac, is 8.3AacI use a Talema toroidal for two channels,
# Power rating: 300VA, 10A
# Secondary: 2x18V
The maximum continuous DC current that can be drawn from this transformer is ~4.2Adc. It will run hot if worked this hard.
It is often recommended that the maximum continuous current should be around 50% of the transformer rating. This is 2.1Adc.
That is 1.05Adc of bias and quiescent current for each of the two channels. No wonder it is running hot on the exterior surface.
How much hotter will it be when inside the chassis/case when the heatsinks are running at 50degC?
How much hotter is the core of the transformer if the surface is already hot?
That 300VA transformer might run one channel and stay within maximum operating temperature. But that is just a guess, I don't know how hot the ambient will be inside your chassis and I don't know how hot the core is relative to the surface temp.
Last edited:
If wont stick much more to this issue, but just note that if I should do any kind of smoothing, I would start with removing ALL the hard anodising with a hand sanding machine, BEFORE doing any attempts to the smoothing
I just use a drill press and piece of wood bored out with a hole saw. I drill a hole in the center (or near center) and then insert a screw that has a smooth shank up near the head. Cut the head off the screw and mount sand paper to the other side of the cylinder of wood using double sided foam tape. Insert into the drill press and then bring it down onto sand paper to ensure it's flat (perpendicular) and then sand the aluminum; clearing often. I start out at 180 grit and then go to 320-350 wet dry. This will give a mirror like finish that is very flat as long as you don't apply too much pressure in the sanding process. I have also used this process with a hand drill that worked quite well when the drill was held in a fixed position and the heat sink brought in at perpendicular. The diameter has to be large enough to accommodate the device you are going to mount. You could also try using the bottom of the sanding cylinders sold at Harbor Freight.
I have tried using an endmill on a drill press at the highest speed using an x-y axis vice; disastrous! The finish was not very good and I got some bit or work deflection that lead to grooves. I think the use of an endmill is best left to a real machining station and not a drill press.
I have tried using an endmill on a drill press at the highest speed using an x-y axis vice; disastrous! The finish was not very good and I got some bit or work deflection that lead to grooves. I think the use of an endmill is best left to a real machining station and not a drill press.
While I have lapped sinks on a granite surface plate using compound which will make it flat to a few ten thousandths, I think that it is a point less effort (other than it looks coool😀).
if you place a sheet of 320 or 400 grit silicon carbide paper on your cast iron tablesaw top and lapp it that is plenty good.
if you place a sheet of 320 or 400 grit silicon carbide paper on your cast iron tablesaw top and lapp it that is plenty good.
Hi,
I finally reacquired the upgrade bug and built a new F5 module using sj201/sk1530. The first channel went smoothly, but on the second only R12 would bias up with P2. No amount of turning on P1 had any effect on R11 or DC offset (+21volts!). Finally replaced the Toshiba's with FQA's and it works now.
Obviously one of the mosfets is DOA. Can someone please tell me which one?
I finally reacquired the upgrade bug and built a new F5 module using sj201/sk1530. The first channel went smoothly, but on the second only R12 would bias up with P2. No amount of turning on P1 had any effect on R11 or DC offset (+21volts!). Finally replaced the Toshiba's with FQA's and it works now.
Obviously one of the mosfets is DOA. Can someone please tell me which one?
I think amp-guy's approach is closer to what i would suggest, if asked.
Making the surface shiny does not mean that it is flat.
_-_-bear
Making the surface shiny does not mean that it is flat.
_-_-bear
perfect opportunity to replace crystal ball with little mosfet testing setup
![]()
Hey Zen Mod,
Who uses a crystal ball? Tea leaves only!😀
Certainly a mosfet/jFet testing setup is on my to do list after DAC, open-baffle speaker, battery-powered F5 🙂
I'm guessing if I have +DC offset then the N-channel is toast? Hoping to save a swap-out of the working channel mosfet...
Toshiba 2SJ201 & 2SK1530
My batches of Toshiba 2SJ201 and 2SK1530 have been run through my Tektronix curve tracer as of a few minutes ago.
(I thought someone was still relatively current on offering up the same devices in Swap, but I didn't find the thread using the search...?)
Anyhow, I will have a fair number for sale.
That's the good news.
The not as good news is that the two batches did not "group" in a manner where the distribution curves overlapped as much as I had hoped. This means that there are not very many dead on device matches to be had out of the batch. There are a fair number of Pch matches and Nch matches - but they will bias up at different points. That won't be a huge problem as already mentioned in the thread.
This means that my price will of necessity be closer to cost except for the sets (if any that I can sell) that are matched Nch to Pch.
Please wait for my post in Swap Meet, I'll list the available devices and prices early in the week, before jumping in, thanks.
Anyone with a number of J201s that when static bias tested need LESS voltage to reach 1.3amps OR with a number of K1530 that need MORE voltage to reach 1.3amps should contact me for a swap out to get matched sets out of it. PM me if you have a reasonable (4-12 or more) number of "orphans" to offer up...
_-_-bear
My batches of Toshiba 2SJ201 and 2SK1530 have been run through my Tektronix curve tracer as of a few minutes ago.
(I thought someone was still relatively current on offering up the same devices in Swap, but I didn't find the thread using the search...?)
Anyhow, I will have a fair number for sale.
That's the good news.
The not as good news is that the two batches did not "group" in a manner where the distribution curves overlapped as much as I had hoped. This means that there are not very many dead on device matches to be had out of the batch. There are a fair number of Pch matches and Nch matches - but they will bias up at different points. That won't be a huge problem as already mentioned in the thread.
This means that my price will of necessity be closer to cost except for the sets (if any that I can sell) that are matched Nch to Pch.
Please wait for my post in Swap Meet, I'll list the available devices and prices early in the week, before jumping in, thanks.
Anyone with a number of J201s that when static bias tested need LESS voltage to reach 1.3amps OR with a number of K1530 that need MORE voltage to reach 1.3amps should contact me for a swap out to get matched sets out of it. PM me if you have a reasonable (4-12 or more) number of "orphans" to offer up...
_-_-bear
300VA Torroid
Prooptiki,
In addition to Andrew's post above (#7643) the actual current pulses that charge the power caps via the diodes is MUCH higher that that again (often 10 times) so the ability of the Torroid to supply these high instantaneous currents is a spec that's rarely included.
So, the 300VA will "run out of steam" for just one channel, unfortunately.
However, you can reduce these pulses by adding a small series resistor between the diodes and power cap (generally about 0.1 - 0.15R) - sound will be much "smoother" and trannie will run cooler.
This is why the use of a 300VA transformer / channel gives better results.
Prooptiki,
In addition to Andrew's post above (#7643) the actual current pulses that charge the power caps via the diodes is MUCH higher that that again (often 10 times) so the ability of the Torroid to supply these high instantaneous currents is a spec that's rarely included.
So, the 300VA will "run out of steam" for just one channel, unfortunately.
However, you can reduce these pulses by adding a small series resistor between the diodes and power cap (generally about 0.1 - 0.15R) - sound will be much "smoother" and trannie will run cooler.
This is why the use of a 300VA transformer / channel gives better results.
If that is true, I wonder why Mr. Pass is using single 300VA toroid in commercial (stereo) version of F5 which doesn't "run out of steam" even when amp is driving the 2 Ohms load ?....
So, the 300VA will "run out of steam" for just one channel, unfortunately.
.....
These are always included. That is precisely why the transformer manufacturer tells you, me and everyone else to de-rate the transformer to about 60% to 70% when feeding a capacitor input filter.the actual current pulses that charge the power caps via the diodes is MUCH higher that that again (often 10 times) so the ability of the Toroid to supply these high instantaneous currents is a spec that's rarely included.
That's why the 8.3Aac transformer ends up supplying a maximum continuous 4.2Adc after rectifying and smoothing.
yes, thanks, Andrew.
Juma,
So, the 300VA will "run out of steam" for just one channel, unfortunately.
Sorry, mistake - I meant single 300va for both channels
My apologies. In hindsight, "runs out of steam" is a poor choice of words - obviously, Nelson's amps work just fine with single 300VA units - my production F3 is also fitted with a single 300va transformer, block bridges, etc and it works quite well - commercial products have different criteria to DIY.
Now, when DIY, a 300VA / channel for the F5 made a significant change/improvement in the amp's performance that is worth the extra initial design/investment (IMHO) - the same thing applies to the whole power supply, diodes, caps, wire, etc.
For many people, the F5 is the last amp they'll ever build.
Suggestions & observations, not criticism.
Juma,
So, the 300VA will "run out of steam" for just one channel, unfortunately.
Sorry, mistake - I meant single 300va for both channels
My apologies. In hindsight, "runs out of steam" is a poor choice of words - obviously, Nelson's amps work just fine with single 300VA units - my production F3 is also fitted with a single 300va transformer, block bridges, etc and it works quite well - commercial products have different criteria to DIY.
Now, when DIY, a 300VA / channel for the F5 made a significant change/improvement in the amp's performance that is worth the extra initial design/investment (IMHO) - the same thing applies to the whole power supply, diodes, caps, wire, etc.
For many people, the F5 is the last amp they'll ever build.
Suggestions & observations, not criticism.
hello everybody,
I've got 8 of those cool kendeil 4700uF capacitors. What would be the consequences of using only about 20mF per rail instead of 40mF? The reason I ask is I cant get any more of them and I really want them sitting in my PSU.
I've got 8 of those cool kendeil 4700uF capacitors. What would be the consequences of using only about 20mF per rail instead of 40mF? The reason I ask is I cant get any more of them and I really want them sitting in my PSU.
I have rifa peh169 caps 63v 10000uf x10 pieces.I will parallel them.They are big caps but very high quality.What do you think about it?Are they much more for F5 amp?
- Home
- Amplifiers
- Pass Labs
- F5 power amplifier