LCLC filters for Steve's amp

that colorado rocky mountain high.
Post Reply
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

LCLC filters for Steve's amp

Post by Johnny »

Hi Dave,

As you know Steve is building an amp using Telefunken RS241's in the driver stage and 75TL's in the output stage. My amp uses the same tubes.
In my amp I use one power supply for both RS241's (280V B+)and a power supply for both 75TL's (500V B+). Steve's amp is going to have a separate supply for each tube. In my amp, I used WE 354A's for rectification for the 75TL supply and WE 287A's for rectification for the RS241's. Here's the filtering for each supply I used.

LCLC for RS241 supply First L = amophous core, 5.5 H, dcr=43 ohms, 100ma; C=20uF, Second L= Nickel core, 5.7H, dcr=70 ohms, 100ma; C=20uF

LCLC for 75TL supply First L = amophous core, 4.1 H, dcr=66 ohms, 250ma; C=20uF, Second L= Nickel core, 5.8H, dcr=30 ohms, 250ma; C=20uF


For Steve's amp, I would like to use the same filters. However, since he has separate supplies for each tube, the current rating given above can be half. So for the chokes for the RS241 supply, only a 50ma rating is needed and for the chokes for the 75TL supply, only a 125ma rating is needed.

Having a ideas,

Johnny
dave slagle
Posts: 2099
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

First questions... are you really running 50ma for each RS241, and 125ma for each RS241?

what will the final B+ for each tube be?

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

First questions... are you really running 50ma for each RS241, and 125ma for each RS241?
No, each 75TL draws 50ma and each RS241 draws 20ma.


what will the final B+ for each tube be?
The B+ for the 75TLs is 500V and the B+ for the RS241s is 280V

John
dave slagle
Posts: 2099
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Johnny wrote:
each 75TL draws 50ma and each RS241 draws 20ma.

The B+ for the 75TLs is 500V and the B+ for the RS241s is 280V
OK we have a starting point... sort of :-)

i think we should draw upon some of our previous experiences and try to make a few informed decisions on the design of these chokes since we have full control over how they will be used.

the burning question in the back of my mind is what the relationship is between inductance and linearity in a PS choke compared to a plate choke. Your previous tests have suggested that linearity (a larger gap) may be more important to the final sound than the largest possible inductance available. So, I have to wonder if it's possible that the same holds true for PS chokes. My first guess is NO, but previous experience has taught me that everything i have liked in a ps choke has also held true for plate chokes and vice-versa. This makes me sincerely wonder what sonic benefits (if any) could be found from tweeking the gap in a PS choke.

I would use the same basic methodology as you did for your previous plate choke tests which is to alter the gap until you find the extremes. I would think for PS choke usage, this would be the point of unacceptable hum on one side and sluggish sound on the other. This also opens up the tradeoff between the values of L and C in the filter which makes the whole problem a bit more complex to nail down. In this situation when you hit the limit of "bad sound" be it hum or whatever, you can possibly correct it by altering the other (capacitive) leg of the filter. If you do end up doing these tests, i'll try to find someone who will handle the simulation end of the picture to see if we can connect any of this together.

since James probably wants to build the PS with given parts and the aesthetics will prevent easy swaps, maybe we should tweek the gap of a known existing similar PS to give us some guidance. It seems that if we could try a few gap sizes in your amp we might get some ideas for what we should do for steve.

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

I won't be able to do it until August. Steve wants to have his amp finished by the middle of September so I don't think we can wait until then. I think we should base the design of the chokes on what I have already done.


Johnny
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

In my last response I should have added that I'll give the "gapping" of the chokes in my power supply a try some time after July. I'm curious about it too but it won't be in time for Steve's amp.


Johnny
dave slagle
Posts: 2099
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

for the amorphous chokes, it is important to design for the current that is going to be drawn to keep operation as quiet as possible, so i'll go with the DCR's you suggested and alter the gap to assure singe quadrant operation (Bdc > Bac) chances are this will mean that the inductance will be higher than what you had, but with half the current drawn you will need double the inductance so it should all work out nicely.

For the nickel chokes, i think 80% nickel in both cases will be the proper choice. since you have 1/2 the current, we can go to the high nickel and end up at about the same place.

these two options would be the safest that i can think of based on previous experience.

let me know what you think.

dave
Guest

Post by Guest »

Hi Dave,

I like your suggestions. Let's go with it. Thanks again for your help.


John
dave slagle
Posts: 2099
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

here are the nubers i'll shoot for

RS241 280V 20ma (Lcrit 14hy)

Amorphous 30R 18hy 20ma.
80% nickel 34R 5hy 20ma

75TL 500V 50ma (L crit 10hy)

Amorphous 30R 11hy 50ma
80% nickel 62R 6hy 50ma

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

I like the numbers. As soon as they are finished ship them to Steve.

Thanks,

Johnny
dave slagle
Posts: 2099
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

the chokes are done.

I ran into some interesting issues when measuring them at the levels they will be used at. The numbers didn't make sense, i'm sure it was a measurement issue and i'll eventually get to the bottom of it. What i ended up doing was gapping them for best performance in the actual circuit they will be used in. (a LCLC supply)

In each case i simply fed a B+ transformer with a variac and built the supply with 30 mic caps. (sorry no 20's on hand) Then i adjusted the gaps for minimum ripple after each stage. For the amorphous chokes i made sure the filter was behaving as a pure choke input and interestingly enough, as I increased the gap the ripple would go down but when the gap got too big, the ripple would keep going down yet the B+ would start creeping up. It makes sense that if the B+ goes up you are no longer pure choke input which explains why the ripple would continue to go down since a "quasi" cap input filter was starting to appear.

in both cases (500v 50ma and 280V 20ma) for the amorphous choke i increased the gap to the point where the B+ started to creep up and then left the gap at the last "full choke input point"

for the 20ma choke it was .0015 (gold) DCR is 28R
for the 50ma choke it was .003 (green) DCR is 40R

a sort of interesting side note involved the second 20ma amorphous choke. with the .003 gap it was giving me 1/2 the ripple of the other one. I played and played with the clamping tightness but it stayed the same. Then came the SNAP where i inadvertantly shorted the output cap and it suddenly behaved similarly to the other. If i had to guess, the large DC current pulse form the temporary short put a huge DC force on the core seating it tightly together. In its prior state, the gap was just a bit larger causing it to behave as the quasi cap input which fits nicely with the rest of the observations.

for the most part the chokes are quiet. they will need isolation mounting since they will shake anything they are securely attached to but floating in air they just have a dull hummmmm.

I did play a bit with adding a small cap (.25u) at the input. Interestingly enough this eliminates the buzz you would get from improper mounting, (they were sititng on an aluminum chassis and the buzz was just vibration) but did little to the underlying hum. the .25 cap also did little to the output voltage, it only added 4V to the 500V supply BUT it also cut the ripple at the output in half!

Several well respected people have praised the merits on a small "snubber" cap at the input of a C/I filter, so maybe adding the option of a switchable cap at the input of the amp would be a good move. A simple center off DPDT switch would allow for two different caps to be put in circuit (switched when the amp is off) The selection of the cap values can allow for a number of different "tweeks" for the future. A .1u and a .25u could be used to see the differences in "snubber" size. or something larger could be added like a .5u and a 1u to bump the B+ up a bit. The only suggestion i have is be sure they are oil caps, and there is the ability to swap new values and sizes in and out easily. Personally i have no problem with caps up to about 1uf at the input even though it does enter the grey area of cap input. Essentially it's just a selection of compromise you need to determine yourself.

Onto the 80% nickel second chokes.

In both cases i increased the gap until the ripple went up. for any given current the L vs I plots for two different gaps will cross at some point. I simply found the two gaps that crossed at around the current to be used and chose the larger gap.

for the 20ma choke it was .002 (amber) 30R DCR
for the 50ma choke it was .006 (tan+red) 70R DCR

dave
Raj Gupta
Posts: 37
Joined: Fri Feb 11, 2005 2:19 am

Post by Raj Gupta »

Then came the SNAP where i inadvertantly shorted the output cap and it suddenly behaved similarly to the other. If i had to guess, the large DC current pulse form the temporary short put a huge DC force on the core seating it tightly together.
That's pretty interesting. I bet you'll try that trick again, too, won't you?

-j
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

LCLC filter

Post by Johnny »

Hi Dave,

Great post!
I did play a bit with adding a small cap (.25u) at the input. Interestingly enough this eliminates the buzz you would get from improper mounting, (they were sititng on an aluminum chassis and the buzz was just vibration) but did little to the underlying hum. the .25 cap also did little to the output voltage, it only added 4V to the 500V supply BUT it also cut the ripple at the output in half!
When James built my power supply he used the snubber to fine tune the voltage we wanted. It works great.


Johnny
Post Reply