### someone listed on 'amp's.. think it was Johm Lyles... about
15 x diff xfmr makers in the USA. Someone got a cheap quote
from 'magcapengineering' xfmrs.... about $100.00 less than the
Dahl.... then we he got it.... found that the DC secondary
resistance was 65 ohms !! [ 5-10 ohms is more like it.] Use
a large C input filter... and the V reg sucked... the high peak
currents will cook the cheaper xfmr's. Rich claims the Dahl's
will do about 4 x their rating on ssb.
No, the one in mention was Hal Mandel W4HBM. He had Dahl design a
transformer for him, and I can't remember if it was a anode or a
heater, but in two tries, Dahl got it wrong, both transformers went
poof for the rated load. I'm not sure, but afterwards, He got his
money back, and I think had the same transformer company in Kentucky
wind one that worked perfectly right from the box. The windings from
Dahl were wound with too small a wire both times if I recall. I
remember helping him troubleshoot it on amps, and the first thing I
said was the transformer wasn't right, and he found that the case
after checking it.
### Hope it's not Hammond. Hammond makes superb stuff....
however.. they don't have a clue as to how to build Hypersil
anything... just standard EI stuff. Hammomnd thinks the way to
build a plate xfmr is to center tap it. All their plate xfmrs were
known as 'high reactance types'... all had center taps... and all
were made for tube rectifiers... and choke input.
No, they're not Hammond, they're a large company named MCI that
specializes in all types.
Hipersil is only a trademark held by Westinghouse which means high
permeability silicon steel. Hipersil's equivelants are M-6 in EI
cores, Hypertran, and Microsil. Microsil is used by Magnetic Metals.
Hypertran is Armco Steels (AK Steel) brand name for the steel. It is a
cold rolled, grain oriented steel with a silicon content of around 3-
1/2%. Westinghouse actually sold out years ago on the steel used in
the manufacturer. One can build an EI core with M-6 that weighs just
15% more than a C-core. The only savings of the C-core would be the
rounded over corners in weight.
### Dahl and some engineers do it themselves. IF u just take his
standard 45-65-127-253-440 lb cores.... and specify the sec taps u
want... they always turn out great. The primarys are all wound
the same with the +10 0 -10.... 208....240 set up.
How can they select so many voltages for different amps with one core
when the whole design considers the incoming / outgoing voltages and
currents to figure the core size? My guess was, when Hal and I spoke,
that Dahl was taking a few cores and trying to make them work in a
bunch of other designs. You can end up finding out that after it's
started, the size wire you wan't wont fit in the window, or there's
not enough room for the number of turns, and then drop it down to make
it fit. That's not good engineering practice, and the efficiency
wouldn't be good over it. I noticed myself that Dahl only uses a few
sizes of C-cores, and used the same ones in many different amp-brand
replacements. They're all listed in his Ham catalog on their website.
There's probably 1000+ different stock C-core sizes available from
Magnetic Metals.
### Primary KVA = DC input x 1.22 That allows 11% for power
factor... and 10% for core losses. His CCS ratings are based
on a 50 deg C temp rise over ambient. I have seen several...
almost all vault transformers where the temp rise spec is sky
high... typ as much as 155 deg C.[over ambient] Why they run
em so hot is beyond me. Power co's do the same thing with
pole pigs... they run em hot. Ever wonder how a 50 KVA xfmr
can run 16 x homes...all with "200 A" service ?? There is
hardly any oil in em either.... the oil is just to transfer heat
to the outside case [I suspect the cases are aluminium..never
seen rust on em]... and keep moisture out during storage.
A transformer running into a full wave bridge, with capacitor input,
needs about 1.6 to 1.8 times more rms current than the DC supply
current for the load. Most HV supplies require about 1.6 as 1.8 is if
regulators are used afterwards. That has a lot to do with transformer
failures I think when one orders them. See Hammond and Thordarson
transformer engineering portions of their websites, and a number of
others that show the same formulas. You can get by with a little less
for SSB only operation as the book "Reference Data for Radio
Engineers" gives 1.45 times the DC current.
Ref:
These days, they're using met-glass which has a high perm, with a
higher flux density than Hipersil, plus they're immersed in oil. Once
a pole pig is out of the oil, it's kva rating drops drastically. The
heat being carried away is the reason why they can be ran where
they're at. Without oil, at the same load, the wire would cook. Some
air cooled types can be pretty ran hot, at about 115 deg C, and up to
17 kilogaiis flux density if a high-temp wire is used. Though the
magnetizing current rises rapidly above 15 kg and the losses really go
up. If one used formvar or polyimide coated wire at those
temperatures, the transformer would smoke in no time.
### How do u propose to get 3 kva output.... with only 1.5 kva
input ??? U need loads of square/rectangular wire to wind
em.... no round wire is used in any dahl's...other wise u end up
with too many air gaps between turns....and more chance of
corona.
When I say 3 kVA or 1.5 kVA input, I mean the power of the amps there
for. I will simply supply a 3 kVA transformer. The customer needs to
figure the size they need by the math, or have me to do it for them.
Naw, round wire has been used for years and years with no problems.
It's more like he uses square because that's all he wants to stock for
certain sizes. Square wire will arc over too if the insulation used
isn't correct. The insulation thickness in mils, and type is what
determines if it will arc between turns, but here's the kicker,
there's no where near the full voltage between turns, nor between
layers. Matter of fact, there's no where near the voltage between
turns or the layers to puncture standard insulation. If it was, either
he had over 1kv per layer at the least, or between turns, and that
just simply can't be the case. At one tun per volt, if we had 5000V,
and 5000 turns, that's 1 volt per turn. We'd probably have at least 10
layers or more, but at 10 layers, that's 500 volts per layer. I doubt
one could get 500 turns in a layer, most likely 20 layers or more
total. The thing is, the larger the core, the turns per volt drops,
and at 1 turn per volt, that equals about 1 kVA. If he said that the
wires insulation has to match the full output voltage, he needs some
more schooling, or he's yanking your chain.
### Back in the 'old days' pumping dahl's with silicone
varnish glop was an option.... and several of the ones that didn't
pay for the option....and stored em in damp locations... had em
short out, etc. Corona is bad news.
I agree, if they're going to be where it's damp, it should be dunked
in varnish. Actually, the only ones I never did varnish was the ones
in school as they had to be re-wound again by another student.
I'd be interested in these other xfmr builder's specs.
Ten-Tecs' supplier is;
MCI Limited:
Others;
Galaxy Transformer:
and
Ed Dennis, Heritage Transformer:
Heritage Transformer Co, Inc.
13483 Litchfield Rd.
Eastview, KY 42732.
270-862-9877.
e-mail hertiran@...
I don't have a URL for them anymore, and a search didn't show a
website. They used to have one, maybe someone else knows it.
Later......Jim VE7RF
Best,
Will
