Passlabs Amplifier Projects
Having been made aware that the service manuals, including circuit diagrams, for the Aleph Series amplifiers had been released on the PassLabs website, I started looking into the possibility of building one. Initially I had planned on upgrading my A40 to an Aleph5, given the supply rails were the same, however after some research I decided to make the Aleph4.
I built the amplifier successfully and details were posted on the PassLabs site. However, subsequent review of the amp by a number of interested people identified some inconsistencies within the circuit and accompanying description in the service manual. Most of these were minor, had no significant impact and were easily rectified.
Nelson has since altered the circuit diagram and the current manual is correct. The most significant change was the number of output devices was initially listed as 'X3', where it now reads 'X6'.
The effect of this upon the amp I built was to run insufficient bias, i.e. with 1.5ohm source resistors @ 0.5V, the amp was running a total of about 1amp bias, c.f. the 3 amps intended. This was easily corrected by reducing the FET source resistors to 0.68ohm with the dV across these being about 0.6V, as suggested by Nelson.
This sequence of events may actually prove beneficial for some would-be-builders. Clearly, the commercial incarnation of the Aleph4 was built in a "bullet-proof" design frame, and quite understandably so!
Judging by e-mail correspondence I have received, the number, cost and matching of FETs is a major hurdle for many. A total of 24 versus 12 devices will have a significant impact on cost. DIY enthusiasts may choose to make some compromise at this point.
Using 3 devices per bank, my amp runs with equilibrium heatsink temperatures at around 65C. This is hot! This temperature would likely be identical with 24 devices total, since the total power dissipation would be unaltered. However, there would be a difference in watts dissipation per device. Given there is a thermal resistance from the FET junction to case, then from the case to the heatsink, the actual junction temp will be above that of the heatsink by a given amount. With fewer devices and more watts per device, so the junction temp will be raised further with the net effect that the longevity of the devices will be reduced.
How significant is this? OK, from the datasheet for the IRFP244s which I used, the rated continuous current is 15A and power dissipation 150W. The increase in current bias from 0.5 to 1A per device is well within tolerance and each device will be dissipating <50W, so we have a safety factor of > 3x. What may become significant is FET junction temperature; this is rated from -55 to 150°C. Looking at the datasheet, junction-to-case thermal R is nominally zero, but in a worst case may be 0.8°C/W. We have to add to this the resistance from case to sink, which may be as high as 0.5°C/W. If we take a worst case power dissipation of 50W, then the junction temp will be 65°C above the heatsink temp, or 130°C in my case.
These figures are identical for the IRFP240, the only difference being a different VA combination for the same power handling (150W)
In reality the figure will be less than this because we have assumed the worst in all cases. However, the take-home message must be, if you decide to build the amp with 3 devices/bank instead of 6, then you must ensure there is adequate heatsink, such that the idle temp of the amp is preferably no greater than 60°C.
If I were starting from scratch I would use the original configuration of X6 devices and 1.5ohm, but then the choice is yours! My amp has done 20+ hours now and works fine. I cannot see the justification for a complete rebuild (and neither could NP for that matter !).
The circuit diagram following is the corrected version. I have included both 3 and 6-FET mounting PCBs in the PDF file. The circuit PCB provided can be used for either the Aleph 1.2, 4 or 5, simply install 4 or 6 resistors as required in the output section.
The basic format for your amp will likely vary from mine. I have never seen the sort of heatsink I have used commercially available and I would be frightened to guess the cost - they weight a ton! Use a little imagination and forward any results to the Webmaster for the projects gallery.
The amp sounds great. I would encourage anyone reading this file to take the plunge and build either the Aleph4 or 5.
I am happy to be e-mailed with questions, this has ultimately proved to be to my own benefit (large grin!).