[hsusmann]

Well, I thought I should start a new thread on this Philco 90 I picked up the other evening with no speaker assembly.

If I can't find a compatible field coil speaker with output transformer I may
just put in a 3K resistor for the coil and then come up with some
universal output transformer.

In simulating the PS circuit in my
Tina Pro simulator it looks like if I just bump up the output cap
value a bit I'll get good low ripple. I'd go with 10 uF max input cap
so as not to stress the 80 Rect. tube too much.

It also appears that I can do even better on ripple if I split the
3K resistor into two 1.5K ones with an extra C in between. May
try that also. Has anyone done this before?

Does anyone have an idea where the best place to get such
a generic audio output transformer might be?

This is the early 90 with 2 x type 45 push pull output tubes.
Philco speaker web page lists it as a K-2 variety speaker assembly
with xfrmr Zpri = 6700 Ohms, Rpri=500 Ohms.

Will post pictures soon I hope!

Herb S.
Ithaca NY

[TA Forbes]

Hello, Herb: just remember that any increase in cap charging in your B+ circuitry is going to result in more current until the caps charge. THEN the benefit of your system to decrease ripple will be same-same.

I experimented with a similar thing with an RCA T10-1, but it has a 5Z3 and consequently can stand the current draw better.

I would be careful with your 80 rectifier.

[hsusmann]

OK, will do. I simulated having a 10 uF cap right on the input to the power supply filter (right onto the 80 filament), then 1.5K then 22 uF to ground, then another 1.5k and then another 22 uF.

The simulator showed very low ripple like 40mv pp.

I suppose I could back off the 22uf's to 10 uf's and increase the
ripple a bit but decrease the overall capacitance that has to charge.

I figure with a total of 3K resistance in there it should limit the current
that the 80 has to supply upon power up.

I should check in my simulator what the peak current is under these
conditions and make sure it is not too high.

The 80 Rect datasheet gives a typical capacitance of 10 uF for a
capacitor input filter so I figure that part is OK. It's the additional
R's with C's that I suppose could be a problem.

Herb S.

[morzh]

80 tube can take up to 32uF direct cap load. Should stay below that.

[hsusmann]

I just checked the 80 Rect datasheet again and it lists max peak transient plate current at 2.2 amps.

In my simulator I get about 890mA peak plate current for the very first cycle of the 120 Hz. So, I think I'm OK. I have made some assumptions about the 80 Rect. model which I think are close, I modeled it as two 1N4007 diodes each in series with 300 ohms. Even if I'm off by a factor of two on the 300 Ohms I'm still OK.

Also, the 1.5K resistors seem to really isolate the later two Caps from the startup transient. Making them 10uF instead of 22uF hardly
changed the first cycle peak current at all.

This makes sense as the dominant factor is charging the input cap
(here 10uF directly from the 80 rectifier output). Initially that cap
looks like a dead short so the current is limited by the impedance
of the 80 rect. plus whatever transformer resistance and leakage inductance there is.

The two C's on the other side of the 1.5Ks don't seem to factor
in at all to that first cycle peak current surge.

Herb S.

[hsusmann]

In further thinking about this problem I realized two main
things that I had overlooked:

1. My simulator was set for an initial phase angle of 0
for the sine wave from the wall socket. Obviously this
is the best case scenario. I changed it to 90 degrees
which should be the worst case scenario and the
peak first cycle current went up to 1.4 amps!

2. However, I think this is all moot as in real life
the tube is cold when you first throw the power switch
so there is no way it's going to be this bad. As the tube
warms up it's conductance will improve so there is a
built-in soft start.

I suppose one could worry about somehow plugging
in an already warmed up set so you really do get full
tube conductance on the first cycle. In that case I think
I'm still almost a factor of two away from a problem.

The 32 uF max capacitance on the 80 rect probably has
more to do with the repetitive current pulses on an ongoing
basis rather than anything to do with first powerup. It then
also should matter what kind of capacitor you have and what
its equiv. series resistance is etc...

Anyone have any thoughts on this interesting problem?

Herb S.

[codefox1]

Why not just use a couple of 10 cent diodes. You can leave the 80 glowing but otherwise disconnected. B+ might need to be tamed some with a chassis bolted 20 watt resistor.

[hsusmann]

Yes, that thought had crossed my mind.

The problem is where do you stop?

I think if I can't find another speaker assembly
I'll go the two 1.5K resistor route with three filter
caps, 10uf, 22uF, and 22uF down the chain.

Now the problem is finding a reasonably priced 8"
speaker. Seems speaker prices are way up from
the last time I looked. Used to be MCM Electronics
carried some $10 type speakers in this size but no
more.

Cheapest I'm finding is $26 now... May have to
cruise the junk stores and see if I can find something
used.

Herb S.
Ithaca NY

[bfrohwein]

Radiodaze has an 8" speaker for 13 bucks. I bought one back in September to put in my 624 that had a messed up VC. It sounds good. it even has mounting holes for an output trans. the Hammond I bought for it fit the holes

http://www.radiodaze.com/product/1095.aspx

Also.. parts express has one for 15.. It was between parts express and Radiodaze.. but I chose radiodaze because of the mounting holes for the OPT and I'd rather support Radiodaze.. I'd be up a creek without them!

[hsusmann]

Wow! good deal. I just ordered two, one for a spare.
These are the kind of prices I remember MCM used to have
but now they are much higher.

I bought a NOS Thordarson audio output transformer off of eBay the other
day that should work find for this radio.

Looks like I'm in business!

Thanks for the tip!

Herb S.
Ithaca NY