Question for the transformer winders.

[retailer]

Does any one on this forum wind their own transformers, what sort of current densities are you working with when sizing the wire.?

In reading about transformer winding (not only in text books but also all over internet) I see suggested wire current density in the order of 800 - 1000 circular mills / amp or approx 1.55 - 2.0 amps / sq mm. My copy of the Radio Amateur Handbook dated 1938 has complete chapter devoted to power supplies and transformer winding - this also suggests current densities of 800 - 1000 circular mills / amp.

Back in the early 80's I had a second job working from home making PA amps for a hire business and this included winding the transformers, they were quite simple - plastic bobbin with only 2 biflar wound secondaries which were connected to give 70v with centre tap I always worked on 3 amps / sq mm - this was suggested to me by the company that supplied the cores, bobbins and wire. I never had a single transformer failure and that was from amps that at times pumped out music flat out for hours on end. Insulation was mylar tape and insulation cut from sheets that had the look and feel of parchment. I have wound transformers for my own use and have at times used up to 5 amps / sq mm, none have ever failed.

Is this figure of 800 - 1000 circular mills / amp simply a carry over from the days when insulation consisted of varnish impregnated paper, pitch and cloth tape. I've dismantled quite a few modern transformers and in measuring the wire size I don't see current densities of 1.55 - 2.0 amps / sq mm being used - more like 2.5 - 4 amps / sq mm.

I did a small experiment - I soldered flying leads onto a short length of .5 sq mm wire and placed it in a cardboard box along with the thermocouple that came with one of my multimeters. This was all wrapped in many layers of newspaper to simulate the inside of a transformer. I then ran 1.5 amps through the wire ( 3 amps / sq mm) and waited for the temperature to rise, 4 hours later the temp had gone up less than 0.5 deg C, it was not until I had set the current density to around 8-9 amps / sq mm that I started notice an increase in the temp.

[Ed_Dinning]

Hi Typical modern transformers with standard insulation are run about 2-3A mm sq. With better insulation it can be run harder, but losses will increase.
Similar reasoning applies to the cores 1Tesla for normal Stalloy, up to 1.5T for unisil GOSS lams.
Transformer design is all about good engineering and understanding all the trade offs. This is essential to be commercially successful. To wind your own transformers, not quite so vital.

Ed

[joebog1]

I use 700 cm/A as a general rule for power transformers. My father taught me how to wind when I was about 10 or 11 and he always said " you will never burn it at that rating".
If I am making output transfomers I use something less ~ 900 to 1000cm, but for guitar use I use about 500 cm. Todays insulation is far superior to what we used to use, even if the modern polyurethane type varnish is solder through. I still try and buy varnish though, but then I am very traditional and conservative when it comes to transformers. Modern toroids are truly horrible for me, so I always design them myself, and have a commercial company wind them to my specifications. There is nothing worse than losing all the skin on your fingers due to the heat. Another thread here discusses this idea too, in asking what is worth keeping today, and will modern stuff be worth repairing. Give me old traditional conservative designs.

Joe

[retailer]

Hello Ed thanks for the reply.
If I'm re purposing a salvaged transformer I connect it up and measure output voltages, after stripping away the laminations I unwind and count turns to give me a turns/volt figure. Once the laminations are cleaned up I reassemble the core with around 100 - 200 turns on it (depending on the core size) and hook up a circuit with a variac where I can measure the voltage across and current through the 200 turn winding. While increasing the voltage around 0.5 -1 volt at a time I note voltage and current readings, the voltage goes up steadily but current stays relatively constant until the core starts to saturate at which point the current starts to rise, this gives me a maximum turns/volt figure and using the universal transformer equation I can get a good idea of the grade of laminations used and then decide on a turns/volt to use for my project. Sometimes I'll plot the voltage and current readings on a graph, this often helps to pick a suitable turns /volt figure. The two turns/volt figures I get do not always agree I feel it's due to the transformer being run partially or even fully sataurated.

Back in the 80's I had to do all transformer calculations with paper and pencil now I have this wonderful thing called Excel, now my spreadseet is setup it's now very easy to play around with turns/volt and the wire sizes I have on hand to get a good fit.

[retailer]

Joe - sounds like you are using lightly lower current densities to me, the other difference being I work in metric (amps per sq mm) and you are using the traditional imperial circular mills per amp. Took me a while to figure out how to convert one to the other so I could compare - not only are they different being metric and imperial but one is area size per amp and the other is amp per area size. I think my calculations are correct here: 700cm/A = 2.25 amps/sq mm, 1000cm/A = 1.57 amps/sq mm and 500cm/A = 3.16 amps/sq mm. I don't have a full range of wire sizes so at times I need to compromise and stretch out the current density closer to 4 amps/sq mm.

With regards to varnish I have a spray can of red coloured polyester varnish called Isonyl and if varnishing then I usually varnish as I go mainly to anchor the turns, I've come across quite a few failed guitar amp output transformers that have been not layer wound but wound on a plastic bobbin with what I call controlled scramble windings, one of the primary halves measures as a partial short but when unwound to count the turns I don't find any burn pinholes or wires welded together. These transformers have been dipped in varnish but it's only a token effort as the varnish has barely penetrated beyond the first few mm. My take on the failure is that the wires vibrate (current flowing through the wire in a magnetic field) and eventually the insulation is worn through causing a short between turns, being wound in a controlled scramble fashion the shorted wires may be 50 or more turns apart.

The older point to point stuff is certainly easier to repair and as you know most of the older audio stuff in now very sought after and collectible. I don't like repairing modern gear and will only do my own or possibly for a family member - some of it is very cheaply made, tracks lift off circuit boards as soon as you look at them and spare parts that mostly come from China are really hit and miss quality.

[joebog1]

I was taught in pounds, shillings and pence, so I still do it that way.

The scramble wound transformers are, as you rightly point out, JUNK!! But I guess you get what you pay for. A "famous" amplifier from the states, now made in China has hard copper flat ribbon cable from the driver board, to the output valve board, and uses 1/4 watt resistors in the phase splitter. They regularly fail!!. To use 1 watt resistors would cost about .000000000000000001 cent more in the quantities they use. WHY?? I have unwound a couple of the above transfomers and found wildly different turns count, either side of the centre tap. I have since designed a 5 section transformer - 3 secondaries and two primaries, on the same sized core, and touch wood I haven't had a failure.

Brings us back to the other current thread, Is it worth repairing? I used to repair those amps, but I don't touch them anymore. You will know when you have the model I am talking about, the chassis is bent into an L shape and is about .6mm thick!!!!!.

If you have a copy of the red brick ( AKA The Radiotron Designers Handbook) you will get a real understanding of transformers of all types. I thought that Langford Smith did an excellent job, and I still resort to the book when I need do "special" designs.

All the best in your endeavours

Joe

[Ed_Dinning]

Hi Retailer, your "take apart and rewind" for transformers was advocated in a WW article many years ago as a good way of repurposing transformers and getting them about right.

I would disagree with Joe about using the imperial calculations; I was brought up in the imperial era; schools then changed to "mock metric" (cgs) and finally at university we went on to MKS. For magnetics I much prefer the MKS system as most figures come out in the 1 to 10 range, which was useful on the slide rules we used in those days!

There is another recent thread where further reasoning is given about the use of MKS for magnetics.

Ed

[retailer]

Sounds like we about the same age, I was bought up with imperial and can still remember grade 4 arithmetic "3 1/2 lbs of flour at tuppence ha'penny a pound" later years of high school science were CGS but radio trade school in 1968 was curiously a mixture of imperial and CGS. Once I landed at the school of sciences at University it was all MKS. A lot of the info on the internet and also text books are from the USA so while learning and picking up information you don't have a choice but be familiar with both.
Even though I haven't had a transformer failure (that I know of) I was still interested in how others did things and in particular wire current densities.

Joe I'm currently doing a scratch built VOX AC15 copy for a relative and used my home made pan brake (with fingers) for the first time to fold up the chassis, looks very professional and so much easier than getting a second rate job after hours of mucking around with bits of timber, clamps and rubber mallets.

[joebog1]

It is a nice looking chassis !!! I get mine bent up for me by a mate who is a plumber and has all the nice toys for chassis bashing. I drill all my holes though. Another couple of cyclones running around so still no paint on my chassis. I now have all the parts though.

Cheers
Joe