Hafler/Keroes 6L6 UL

[sorenj07]

We continue the program with Soren just having received his MASSIVE 60-watt Edcors, which look great. Anyway, the choke (hammond 193k) hasn't arrived yet, but after looking at some stuff in PSUDII, I have a couple more questions about my choke power supply. With a 47uF input cap and 150uF cap after it, this choke with a H of 2.6, DCR of 21 and mA handling of 300mA sees around 1.9A in a big spike that settles down pretty quickly to 300mA (this is with a 300mA load). I'm just wondering whether I should be concerned with this spike. I can't be overloading the choke with the 47uF input, right?

[Tweeker]

Is this current spike just perhaps capacitor charging currents, or is there more going on? Does the voltage come up smoothly? Try using a stepped load in PSUDII.

[sorenj07]

hmm. what would a stepped load do? and what should i use for the two levels? total load should be 300mA, so should I step it at about half, and when should I step it? I really have no clue here.

[Tweeker]

Its a way to look at if the high peak current is just capacitor charging inrush, or something else. The values arent that critical.

[EC8010]

Try stepping the current from 30mA to 300mA and look to see if there's any low frequency oscillation. Set the time scanned to be 3 seconds or so. Then try stepping from 300mA to 30mA. The stepped load is very good for revealing underdamped power supplies. Juggling the value of the first capacitor often helps.

By the way, motor start capacitors are notoriously unreliable. Use motor run.

[sorenj07]

Well I stepped it from 30mA to 300mA after 2 seconds, and still got a spike up to around 1.25A that "rang" above and below 0A with 9 peaks in the first 1/2 second, then it quieted down. The problem is, I already lowered the second cap to 47uF so my CLC was 47uF, 2.6h/21R, 47uF and I'm still having this problem

Another thing I'm considering (although it would take around $250 more that I don't have just yet) is to go with monoblock config. I realize that the load per channel is 130mA and not 150mA btw. In any case, I'm considering using a 5AR4 for full-wave rectification of a 300-0-300V 272JX. It'd still use a CLC filter, but in this case both C's would be two 50uF 500V sections of a chassis-mount can cap (I'm a sucker for those) and the L would be a Hammond 193J, which offers 10H/82R/200mA. The biggest spike I'm seeing with this config is around 700mA, and it's a lot lower than 1.2A, that's for sure... With soft-start enabled, however, seeing as this _is_ a vacuum rectifier, the biggest spike is 170mA, so I think I might be good

[Tweeker]

Try raising the second cap to ~300uF rather than lowering it. Use a thermistor on the primary to cut inrush down.

I think you need to use Hammond 193M (10H, 63R, 300ma) rather than 193J. If your buying all this iron, consider choke input (may need 193Q, 10H, 53R, 500ma $63 from Angela Instruments). You might do something like this stero psu to save money over mono while having good channel separation as so:

........R-C
L-C<
........R-C

The 5AR4 isnt instant on, it should mitigate the startup inrush, should it not? PSUDII assumes the heaters are already hot.

[sorenj07]

Originally posted by Tweeker
Try raising the second cap to ~300uF rather than lowering it. Use a thermistor on the primary to cut inrush down.

I think you need to use Hammond 193M (10H, 63R, 300ma) rather than 193J. If your buying all this iron, consider choke input (may need 193Q, 10H, 53R, 500ma $63 from Angela Instruments). You might do something like this stero psu to save money over mono while having good channel separation as so:

........R-C
L-C<
........R-C

The 5AR4 isnt instant on, it should mitigate the startup inrush, should it not? PSUDII assumes the heaters are already hot.

At this point, I'm fairly certain I'm going to go with dual-mono. This project is probably the only tube amp I'll have until I'm out of college and have a real job, so I don't want to have any regrets. Plus, monoblocks just plain look cool . In any case, this is the latest iteration of a PSU design. I was considering using a CLCLCR with a 350V transformer but honestly the price would go through the roof.

http://img227.imageshack.us/my.php?i...l6l6psutd5.jpg

[Tweeker]

Your buying big new chokes and new transformers, why not dual mono choke input with 279X and 193J, and a pair of 80-100uF poly in oil motor run caps on filtering?

[sorenj07]

the main reason I can't do that is that my PT is only 275V, so best-case scenario, I can get 364V with _minimal_ filtering i.e. a small choke. If I want to really filter the heck out of the power, I need to start with a higher voltage because of the drop. I'm even considering getting a 350V model, using vacuum rectification, a CLC filter, and a final RC filter to bring the voltage down to something workable. This monoblock style with a 272JX feeding a 5AR4 feeding a CLC filter with the 193M and two 50uF LCR can sections is probaly what I'm going to go with. Just gotta wait a few (many) weeks until I can scrounge the cash. It'll be worth it, I suspect.

One last thing though - if I were to stick a 100+uF motor run cap in the CLC filter just described, it'd be best after the choke, right? I do have a Xicon 47uF 450V axial that I could use instead.