[Amps] Alpha 78 rebuild, some answers and more questions

[Will Matney]

Martin,

See below,


> 
> 
> A great response with some good ideas about measuring the airflow. Several
> mentioned the method proposed by G3SEK and I see there were a couple of
> Radcomn articles in 97 about airflow, unfortunately I disposed of my pre 98
> magazines when I moved here so have not been able to read them.
> 
> I used the formula 1.76 * (Pdiss / delta T) suggested by Dirk, ON4ADZ, this
> gave an approximation for airflow in cfm of 21. Comparing this with the bin
> bag method suggested by several, I saw a 3.5 cu ft (approx) bin bag fill in
> some 9 seconds averaged over 5 tries. I make that equal about 23.3 cfm.
> Given a healthy dose of realism +/- 10% on both figures suggest this might
> not be far from the truth so I'm assuming I have something just over 20cfm
> of airflow through the tubes, this is with both fans operating and more than
> the Eimac stated 8.6 cfm with 400 watts dissipation
> 
> I did try wiring the internal fan directly across the 120 VAC winding of the
> transformer but it certainly makes a lot more noise for very little change
> in airflow. The external fan, a 4.5"  diameter 220 VAC bolted to the power
> supply air inlet with a homemade foam rubber gasket adds less noise and
> gives a much greater airflow.
> 
> I have re-soldered the coil tap with a big 200+ watt iron. Seems to be fine
> so far and I have seen about 1.2Kw out on 15 and 10.
> 

I doubt you'll have any problems. The only ones that do are the ones who switch from a large 10 meter coil to a smaller solid wire wire coil for the lower bands. On these, the wire size transition is too great and the turns closest to the 10 meter coil can get hot. On the ones that do, you can readily tell it, those turns will take on a dark blue color.

> Starting to feel that amp number 1 is okay and working as it should with
> enough air movement so things don't melt as they tend to do here. Attention
> is turning to amplifier number 2, having robbed it for its multimeter and
> one of the tubes what I have right now is essentially a chassis with most of
> the parts. My thinking is that much of the hard work is done, just the
> remaining 50% to get it back together as some sort of amplifier. There are
> some bits missing, the aforementioned multimeter which I know I can replace,
> some broken switches which I know can be sorted and strangely 3 broken
> ceramic standoffs, the two big ones that are at each end of the main PA coil
> and the small one near the choke, not sure about replacements for these yet.
> The front panel is pretty rough but that’s only cosmetic.


The main manufacturer for those ceramic standoffs was the Herman H Smith company, now named Abbatron Products. They list them on their website but the pics don't show for some reason. They have both types, either the straight standoff, or the feedthrough type. There is one other company that makes these I know of, but I cant seem to find them on the net, or remember their name. They carry wire and other parts too. Someone on here may remember their name. If it's a HH Smith part, it will have a part number either molded into it or printed on it somewhere. Their website is at;

http://www.abbatron.com/products.php

Look under HV Insulators and Ceramic Standoffs. They can help you with the size and direct you to the closest distributor. One hint, those parts were made to industry standards where one companies would fit another. You may even be able to get 1-2 as free samples.

> 
> The big question has to be tubes, and I know this has come up on here before
> but I can't believe its not fixable. 3 more 8874's don't really seem
> sensible unless somebody has a pile of good ones for $50 a piece. The
> modification to use a pair of 3CX800's seems equally expensive both now and
> going forward. So what to do. I have seen suggestions of using russian tubes
> but do not know of any actual work done to use these on this style of amp.

There's one Ruskie tube, I can't remember the part number, which has a removable cast radiator I think would fit the bill. Tom at Tom's Tubes can help you there.

> 
> In order to cover as many possibilities as possible here is what I think
> could easily be done:
> 
> Remove the plate switching relay to give room for an additional transformer.
> I don't feel the CW and SSB selection is a real necessity.

If the transformer is good, why would you need an extra unless going up in tube size, power, or changing to a tetrode. It's best to leave an antenna relay inside the chassis where it's shielded, receives cooling, and is close to the coax jacks. It's best to have the switch legs on the relay connected directly to the jacks by a short flexible wire. I've seen a lot of problems mounting a relay away from the jacks and trying to run coax to it. In RF, it's best to keep all leads as short as possible. Generally, the relay and jacks will be right at the tubes. Then the input to the tubes has a short connection. The only one with a long connection of coax is generally the output from the load C to the relay. Some are not like this like the Heathkit SB-220 which has the input tune network up front and the relay mounted in the middle of the chassis. I don't like doing this personally. A good layout will have the input tune networks and input bandswitch right at the tubes in the rear. Then as above, the only long length to connect would be the load C to the antenna relay. This keeps all the connections as short as possible to keep away from stray RF.


> 
> Re-do the rectifier and filter. I have capacitors half the size of the
> originals which would either allow more space for other psu hardware or
> allow more capacitors with a voltage doubler depending on tube.
> 
> Replace the tube plenum box with one for the tube or tubes of choice.
> 
> A single large triode would seem the easiest but I don't think there is
> enough room for the likes of an 8877.


Again, ask Tom at Tom's Tubes on this. He can get you set up with a cheap ruskie alternative that will perform good. He also has the sockets for the one tube I mentioned above.

> 
> There has to be an alternative short of fabricating a whole new chassis so
> what would seem to be the way to go?
> 
> 
> 73
> Martin HS0ZED
> 
> --

Best,

Will

-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm