Recently I have been moving away from vacuum tube power amplifiers and more towards solid state amps capable of driving a pair of Electrostatics, namely an older pair of Martin Logan made around 1995. These speakers have newish panels and rebuilt cross-overs with brand new film caps. However these speakers are a difficult load and quite current hungry. I have also tried these speakers with other power amps, namely a Korsun / Dussun v6i in power amp mode capable of 400W into 4R, an Accuphase P180, a Crown Pro Audio, a few tube amps and finally this F5 which is my main amp at present. This amp was a scratch build using PCB's bought via a group buy. Case is from Italy and other parts from RS and CPC. I used premium parts when available and later upgraded to Aerovox caps for the PSU.
PCB's and full parts are commonly available on eBay these days if you search world wide if you don't mind a cheaper chinese copy.
"The power supply of the tested amplifier is +/-24 volts and should be rated at 6 amps continuous duty, and more than 10 amps peak per channel. Ordinary unregulated supplies will work fine" (Neslon)
I used 2 x 300VA transformers for the PSU and they only just get warm to touch so are not stressed or running near their limit. I know some have built this amp with 500VA per side or even bigger!! Class A amps are very in-efficient, typically around 20% with about 80% of the energy wasted as heat. So as a quick calculation :-
Lets say the transformers are 90% efficient 300VA = 300W @ 90% = 270W Power per tx = 540W
Power Out = 25W at 20% so 80% Loss = 100W Lost as Heat
Total Power Used Per Side is 20/100 x Pt = 25, so Pt = 125W
We need an Heat-sink that can dissipate 100W heat per side.
Total Power Used ~ 125W per side ~ 250W Total
So we are using just under 50% of the Mains Transformer power capacity. Anyhow, that is how I worked out the requirement. Maybe not exact maths or science but close enough for me and the mains transformers run cool. So using a 1KW or 2x 500VA Transformers is a bit overkill IMHO.
In the images I later removed the protection fuses at the speaker outputs and replaced them with a protection circuit. Fuses create a non linear response and ideally a heavy duty protection relay should be used here. I am not sure how Leak got away with this in their early transistor amps. The protection circuit offers a delay at power-on so the speakers are switched in after 10s. It also disconnects immediately at power off and if it senses DC on the output in a fault condition.
I later removed the protection circuit so there are no contacts in the output audio path.. it sounds vastly better. The circuit has short circuit limiting by design and there is no switch-on-off thumps so the only risk is DC in a fault condition which is rare.
Sound as you would expect from class A is smooth without sounding artificially warm and fat like many tube amps. It has the speed and clean sound of a high end solid state without the grain or haze. Not the biggest and widest sound-stage but ample in a modest room. I have had tube amps with a much wider and bigger sound stage but many tube designs can sound slow, fat and lack control, speed and timing of a good solid state amp on some styles of music. Tubes and Solid State both have their merits.
Considering this amp is only 25W with a current limit at 10A per channel it seems to drive the Martin Logan Aerius ok. It is probably not the perfect power amplifier for these speakers and most people with Logan's seem to use some seriously large and powerful, but very expensive amplifiers. I think in a large room it could be out of its depth on these speakers. I built it because the PCB's became available by chance and it was easy and cost effective to make a high end amplifier for under 1K with the best parts. Thanks to Mr Pass for making his design available to the DIY Audio community and for such a fantastic design.
- It has very wide bandwidth, DC to > 500 KHz.
- No capacitors anywhere in the circuit. (except in the power supply)
- It has a high input impedance – 100 Kohms, and a high damping factor (~60)
- The distortion is very low, between .001% and .005% at 1 watt.
- It’s very quiet, about 60 microvolts or so.
- It will drive a 2 ohm load without burping, and 1 ohm without misbehaving.
Nelson Pass 5/24/08
The top down image with the long green PCB's is the commercial version of the F5 from Nelson Pass for comparison.
Upgraded Pass Labs F5 PSU
A year laster I decided to upgrade the standard build power supply in my Pass Labs F5 with some large screw terminal Aerovox caps from CPC. My original build used 4 x 22,000uF snap-in caps shared between both channels L+R on the PCB which is the recommended minimum. Some have used 2 of these PSU PCB boards, 1 per side which is the next step up. My late friend Dave Hewitt used 4 large caps in his F3 so I eventually followed and bought 4. This provided 47,000uF per cap so more than double the original capacitance and storage. I also by-passed each side of the supply lines with some decent quality 10uF film caps. The new caps are each about the size of a Red Bull can.
Sound and Aerovox Caps
It seems these caps need to run in quite a bit. They behave just like Black Gates... At first listen the amp sounded obviously better. More weighty, better stronger bass, solidity and a wider sound stage. The F5 always sounded a little small compared to a good tube amp with the regular PSU board so this was a very good upgrade even if the caps are not so cheap. However this improvement in sound quickly diminished and the amp became very lean and a little gritty sounding. At this point I was just hoping that these Aerovox caps would improve with time. I am now around 100 to 120 hours burn-in and they have certainly changed again. The sound has returned to the more fleshed out, solid and mellow tone that the amp had originally. I think these really need quite a few 100 hours at least to settle, form or 'burn-in', whatever you want to call it.
I was a little surprised at this and never really considered such caps needing much if any burn-in but not so. These really need run time just to sound right and do get better the longer they are used. Don't be put off and rip them out as you maybe tempted to do during this period, patience is key.
I have compared this amp to other DIY builds and my commercial Yamaha AS1000 and it beats the AS1000 in terms of finese and detail. The AS1000 is a little bloated in comparison. Compared to 300B, 2A3 tube builds it is not far behind in terms of mid range but vastly better in other areas such as bandwidth, timing and speed for example. It really is the best of both tube and solid state worlds.