Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Jul Thu 18, 2019 6:01 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 15 posts ] 
Author Message
 Post subject: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 1:04 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26015
Location: Annapolis, MD
Stop me if you've heard this one....

There are various tests that one might like to do an an OPT. The most common is probably determining the turns ratio. Standard methods include inductance measurements and measuring the transfer function using the output of a Variac. Both are fine at some level, but have issues.

Ideally, we would like to be able to measure the transfer function over a range of frequencies. This requires a general-purpose audio oscillator and a good power output stage and a high-voltage supply.

The new twist: Drive the transformer in reverse using a modern S/S amplifier. With good linearity, frequency response, and low output impedance, this has the potential of minimizing the variable and error sources. I've done a trial run on an OPT I just wound for a Philco 116B. I can measure the turns ratio and any frequency and--in the process-- see the frequency response of the transformer. One puzzle came when i tried to see core saturation. Just monitoring the primary voltage, I was well over 800 volts p-p and did not see any evidence of saturation. Next step will be to repeat the test while monitoring the current into the secondary.

I'm also thinking about setting up an op-amp to measure the ratio between input and output--to make the measurements a bit less tedious---especially if mapping the frequency response.

Looking for ideas as I dig into this....

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 1:39 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 5249
Location: Rochester NY USA
You should probably do the saturation test at 20 Hz - at 60 Hz or more you may hit insulation breakdown first! It can be scaled to other frequencies - will be twice the voltage at 40 Hz, etc. Current waveform should show it more clearly with a low impedance source.

_________________
My web page: https://bit.ly/2rxq4qx


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 1:51 am 
Member

Joined: Sep Tue 15, 2015 1:16 am
Posts: 532
Location: 18424 PA
I would think to see any saturation, you will need the proper load. Chances are 99% of them being used will not get in saturation anyway. If you want proper answers go to diyaudio.com and join the forum there.

I use a USB scope and it does a freq sweep to whatever you set it to. Shows phase shift also. Also much easier to just use a audio signal generator and put 1v on the secondary, measure your primary voltage.


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 12:01 pm 
Member
User avatar

Joined: Jan Mon 18, 2010 2:13 am
Posts: 18498
Location: Dayton Ohio
Interesting ideas.

About all Ive done is connect a small power transformer (12.6 VCT secondary) to a variac and connect to a transformer under test.
Adjust for 1 or 10 volts depending on the transformer and then figure out the turns ratio.

Since you've been winding your own, I can see how you want to know more about their performance :D

-Steve

_________________
Radio Interests
-Zenith
-Sparton
-Pre-War FM
Consoles and floor models, the bigger, the better!


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 3:04 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26015
Location: Annapolis, MD
A subtlety here is that I am not building transformers in the sense that I design and build for an application. I am doing "copycat" RE-builds. However, that is not always possible**. I just finished one transformer where i reduced the number of turns by ~10% to make everything fit (keeping the same turns ratio. I'm motivated to verify that this hasn't caused a performance issue.
As soon as some garden chores are done, I'll go back in the cave and make some more serious measurements.

**take apart a small OPT and see how the primary was meticulously wound in perfect layers, with a very thin piece of (wax?) paper between layers. Even with the most sophisticated equipment, that must have been pretty tedious work. With modern magnet wire insulation, you can "scramble wind" with no insulation between layers. This makes something like an OPT a relatively easy task.

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 23, 2019 5:52 pm 
Member
User avatar

Joined: Jan Mon 18, 2010 2:13 am
Posts: 18498
Location: Dayton Ohio
Yes, I can see that an output transformer would be fairly easy. The primary going down first with the heavier gauge secondary on top.

I knew a guy maybe a few years older than myself who would rewind interstage audio transformers.
Using incredibly fine wire in layers with wax paper between them. He had the patience of Job...

I have a Zenith audio transformer (MS-200 used in several 1933 models) which I want to figure out how to rebuild.
Finding a good one that is intact seems to be near impossible. :cry:
I know that it is a step-down "Driver" type, so maybe it won't be near as difficult as winding a 1:3 type.

I want to know what the actual number of turns is and of what gauge wire, so one of these days I'll tear one apart and see if I can sucessfully disect it.

-Steve

_________________
Radio Interests
-Zenith
-Sparton
-Pre-War FM
Consoles and floor models, the bigger, the better!


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Sun 26, 2019 2:54 am 
Member

Joined: Jun Fri 19, 2009 6:34 pm
Posts: 8534
Location: Long Island
So how are you dealing with the DC bias issues? Any iron core transformer or choke that normally has DC current in at least some of the windings is normally measured on an incremental inductance bridge, which is an adaptation of a standard bridge that can simultaneously pass DC current and AC test signals through the transformer or choke simultaneously.

The problem is that the magnetization curves of iron and steel are not particularly linear. Therefore the presence of DC bias is going to put you on a different part of the curve than testing with AC only. Generally speaking, DC bias reduces the inductance of a transformer or choke, maybe changing its frequency response and linearity, and brings it closer to saturation.

Granted, the standard issue Philco 116B has push-pull output tubes, so the DC magnetization in one half of the primary tends to be canceled by the magnetization of the other half. But that assumes things are perfectly balanced between both sides of the output stage, which they seldom really are. So it usually does make some slight difference. Of course for single ended output stages, DC bias is crucial towards arriving at the correct "transfer function" as you call it.

_________________
"Hell, there are no rules here--we're trying to accomplish something!"

Thomas A. Edison


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Sun 26, 2019 4:44 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 4361
Location: Sunbury, Ohio 43074
I have to agree with loading an output transformer to get any meaningful results about saturation, or any other parameter really. You’d want to test it under real world conditions.

That would hold for both resistive and inductive loads. Possibly even more involved for speakers, where the load actually “fights back” in a sense

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Sun 26, 2019 7:50 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26015
Location: Annapolis, MD
I thought a bit about the bias issue---next test will have maybe 50 - 100mA DC bias on the secondary. In principle, that means one swing of the AC signal will saturate before the other. One issue: The DC supply needs to have an impedance high enough that I can get the total in the ballpark.

One key assumption: Suppose the OPT is 4000:1 impedance ratio. If the attached speaker is 4 ohms, then the transformer input will look like 16Kohms. I'm assuming that one gets equivalent behavior with 2 options:
--A known voltage on the primary, with the secondary terminated in 4 ohms
--A known voltage on the secondary, with the primary terminated in 16Kohms
One source of error in this is the DC resistance of both windings.

Heading for the cave to get some actual data.....

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Mon 27, 2019 12:46 am 
Member

Joined: Jun Fri 19, 2009 6:34 pm
Posts: 8534
Location: Long Island
Correct transformer protocol would be to have both the primary and the secondary terminated into their nominal impedances. If the speaker is four ohms and the plate load is 5,000 ohms (for example), you'd place a four-ohm resistance across the secondary and a 5=k resistance across the primary. That impedance can consist of the source driving the transformer in parallel with the load resistor, so on your secondary side the resistor would be somewhat higher than four ohms in this example.

Also, in a standard output transformer configuration, the DC bias is on the primary side and it would normally represent the average or DC value of plate current at nominal output power. There is no DC on the secondary side, so no need to apply bias there.

_________________
"Hell, there are no rules here--we're trying to accomplish something!"

Thomas A. Edison


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Mon 27, 2019 12:58 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26015
Location: Annapolis, MD
Yes, but....

In the typical tube output stage, we don't "drive the OPT at it's characteristic impedance". The load on the tube(s) is based on empirically-determined compromise between power and distortion. The actual output impedance of the tube circuit can be very different from the recommended load.

Having said that, it is certainly true that things like frequency response are going to depend on the impedance connected to either winding.

Why am I feeling this 50-year flashback to lab classes in college?---especially the part about explaining why your measurements don't match the theory. ( I was a member of the minority that actually cared.....)

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Mon 27, 2019 1:09 am 
Member

Joined: Jun Fri 19, 2009 6:34 pm
Posts: 8534
Location: Long Island
It's called curiosity, and I do all I can to feed mine.

Regarding output transformers, it is true that manufacturers did not always try to match the recommended plate loads of output tubes exactly as they are spelled out in the RCA Tube Manual. A lower impedance would give more output power, but it comes with higher distortion. A higher impedance would give less distortion, but less output power too. And as previously noted, in the real world, the impedance is anything but constant over the frequency range; a speaker might look like a resistive load at one frequency, an inductive load at a another frequency, and a capacitive load at yet a third frequency.

_________________
"Hell, there are no rules here--we're trying to accomplish something!"

Thomas A. Edison


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Mon 27, 2019 12:24 pm 
Member
User avatar

Joined: Jan Mon 18, 2010 2:13 am
Posts: 18498
Location: Dayton Ohio
Chris108 wrote:
in the real world, the impedance is anything but constant over the frequency range; a speaker might look like a resistive load at one frequency, an inductive load at a another frequency, and a capacitive load at yet a third frequency.


But somehow, everything comes out just fine and brings warm music to our ears. :D

-Steve

_________________
Radio Interests
-Zenith
-Sparton
-Pre-War FM
Consoles and floor models, the bigger, the better!


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Mon 27, 2019 12:32 pm 
Member

Joined: Nov Wed 30, 2016 7:35 pm
Posts: 4361
Location: Sunbury, Ohio 43074
It's a good thing our ears aren't linear, huh. Distortion however, is painful. At least to me. While I am not an audiophile (yet) I do appreciate clear undistorted music. Although even that.... an awful lot of people do not hear, don't notice, or don't care.

Output transformers, and their relationship to the tubes behind them and the speakers ahead of them, are a black art. Always were. Just ask the designers at McIntosh. ;-)

_________________
Preserving the history of electronics, one boat anchor at a time! :)
www.bbtvtestequipment.com


Top
 Profile  
 
 Post subject: Re: Testing output transformers--A new method
PostPosted: May Thu 30, 2019 3:00 pm 
Member
User avatar

Joined: Jun Wed 08, 2011 2:33 am
Posts: 8715
Location: Ohio 45177
MacIntosh has gone one step further designing transformer coupled solid state amplifiers, too. I can tolerate low to moderate distortion based on the source and the bandwidth. Like hearing music on a tube radio with 10-100 times the distortion rating of my hi fi gear. What grates on me is noise and hum, a steady state distortion. The grinding rough sound on line noise on my ham radio grates on me and I eventually turn that off. Even a background noise from computer DSL in my AA5 in here that is well below the signal level eventually shows thru and is a source of irritation. I tune in a station lower down the band then, that lacks the background noise. Sort of a white noise hiss versus line noise.

_________________
Reddy Kilowatt says; You smell smoke? Sorry about that!


Top
 Profile  
 
Post New Topic Post Reply  [ 15 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: Google [Bot] and 5 guests



Search for:
Jump to:  




























Privacy Policy :: Powered by phpBB