OT testing - Please help me to refresh memory and understand what I measure

[kagliostro]

This days I'm rearranging the lab a bit

I take measures on OT I've and label it with the result

Among the transformers there is one (Grundig 9054-339.02) that for dimensions seems a 3W

if I measure the turn ratio it has a turn ratio of 10.5

so 10.5 x 10.5 = 110.25  --  110.25 x 8ohm(load) = 882ohm (primary impedance with 8ohm load)

if I measure it using a diffusion line impedance meter (it performs test at 1KHz) with an 8ohm (resistor) load

I obtain 1.8K (primary impedance with 8ohm load)

---

I remember that at different frequency the return can be different, but I'm not able to remember more on this matter

Please can someone refresh my memory ?

Thanks

Franco

[jjasilli]

See:  http://www.radioremembered.org/outimp.htm


Dos this help?

[kagliostro]

Ciao JJasilli, Thanks

Dos this help?

Yes, a begin of an explanation there is

The impedance load seen by the tube and output transformer is not constant. The frequency of the audio signal will vary over a wide range. The inductance in the windings will have a different impedance at different frequencies. At a certain frequency an 8 ohm voice coil may have an impedance of 10 ohms or at low frequencies it my have an impedance of 4 ohms. This varying load impedance is reflected back to the primary, so the tube, and output transformer must work into a varying impedance range.

Is about the inductance of the winding that I've to investigate

Thanks again

Franco

[jjasilli]

This is why 1000Hz is the standard measuring frequency.  Humans hear best at this frequency so full range speakers are designed so their nominal impedance equals their actual impedance at 1000Hz.  It's also an argument for multiple divers for hi-fi.  E.g., a tweeter can be deigned to have an 8 ohm impedance at higher frequency, while a full range speaker might have a 10 or 20 ohm impedance at that point.

[kagliostro]

Thanks JJasilli

This I can understand (near easily), where I'm a bit (a bit is an euphemism) in difficulties is to put in practice use the measures I do

if the turn ratio metod (established with 50Hz voltage) give me around 900ohm impedance and the measure with the impedence meter

at 1KHz give near 2 times impedence, 1800ohm (with same load at secondary), which is the impedance I must consider to be seen

by the tube ?

I think also that there will be difference in the two measures depending on the quality of the OT, is correct to think thar transformers with the same turn ratio can have different inductance and so, at the use, different reflected impedance at 1KHz ?

As usually we use to measure at 50-60Hz the turn ratio (and then compute using a load value), is this the reflected impedance of the primary to consider or ....... 

Franco

[jjasilli]

No, no.  You have to use a low Hz AC voltage to determine the turns ratio, IF you use a volt meter.  A volt meter can't accurately measure AC across the audio range. A volt meter can properly measure 50 or 60 Hz.

[sluckey]

My Fluke measures up to 100KHz. Accuracy is really not important since you are looking for a relative "ratio" type reading.

[kagliostro]

The line impedance meter that I've is a CESVA MPI-3

I connect a resitor on the secondary and the instrument on the primary, select the impedance measure and the range, the result is directly an impedance value, sometime te value are close to the espected impedance, other time they are not, so I was looking for the reasons

(OK, it was made to test speaker lines, not Output Transformers)



------

I agree, relatively impedance, not a perfect measure, but 900 or 1800 isn't so close

Franco

[jjasilli]

Ok, I think I understand better.  Usually, the OT is "considered" a "perfect" and transparent device. So its characteristics are ignored; we pretend they don't exist.  But you want to measure the internal characteristics of the transformer itself.  As a coil, or set of coils, it will have its own internal static DC resistance; inductance & capacitance.  However, it gets more complex with "eddy currents" and other factors I might never have heard of.  These may vary not only with frequency, but also with the amount of current flow. Actually, it may be that the capacitance & inductance values might be constant.  E.g., a specific inductance value will have a more or less pronounced effect upon different specific frequencies - but that's not a change in inductance value.


Anyway, for this quest I think you need to consult textbook references on output transformer design, theory & testing.

[kagliostro]

Thanks

Franco

[jjasilli]

I did some google searching and found little by way of internal OT testing procedures; and nothing that I would rely on, without reference to established sources like known reference books, EXCEPT for 2 things:


1.  It's dangerous!  Because of the significant voltage step down, when test voltages are applied to the secondary, huge voltages appear on the primary which could exceed 1000 - 1600V.  Such voltages are very dangerous for a number of reasons and require special care.


2.  OT's are often spec'd out @ 1W (same for speakers), which may reveal little about actual performance.

[kagliostro]

Again Thanks

Franco

[sluckey]

I just did a simple check on an AO-24158-2 OT that was salvaged from my AO-63 conversion project. The OT was originally driven by a pair of 6GW8s. Hammond used good iron in their Organ amps and this one is no exception.

Here's the test... I used an HP-200CD as a signal source and made all measurements with a Fluke 87-V True RMS meter. I set the sig gen to 1V @ 60Hz connected to the OT secondary and measured voltage across the primary (plate to plate) leads. There was 30V across the primary. I repeated this check for several other frequencies, always ensuring that 1V was fed into the secondary. Here are the results...

60Hz @ 1V on secondary gave 30V across primary.
500Hz @ 1V on secondary gave 30.3V across primary.
1KHz @ 1V on secondary gave 30.2V across primary.
5KHz @ 1V on secondary gave 30.2V across primary.
10KHz @ 1V on secondary gave 30V across primary.

These readings show me two things...

1. The freq response of this OT is flat between 60HZ to 10KHz. (I'm sure there would be more peaks and valleys to plot if I were to use more frequency samples or even a sweep generator or spectrum analyzer, but I'm not trying to sell an OT. Just making a point.

2. Since the voltage ratio remained constant, the turns ratio is constant, and therefore the impedance calculation remains constant at these various frequencies.

My conclusion is that for a given OT with a flat frequency response, the impedance ratio is based only on the turns ratio and is unaffected by the frequency you choose to do the impedance test. (But if you use a cheap transformer with a poor frequency response, all bets are off.)

I think using 60Hz or 50Hz to do the impedance check is a good idea because it allows anyone to be able to do this check. You don't even need a variac, although they do make this test even easier. All you need is a filament transformer (heck, even just a pair of gator clips connected to the filament winding of an amp that happens to be sitting on your bench or kitchen table). Simply connect the 6.3vac across a 500Ω pot and connect the wiper of the pot to the secondary of the OT under test. Set the pot for exactly 1vac across the secondary and measure the voltage across the plate leads. Do the math.

OK, test time... If I connect an 8Ω speaker to my test Hammond transformer, what will be the reflected primary impedance?

[kagliostro]

Many Thanks Steve

There are many interesting things on what you wrote

things that clarify some points and deserve further study

Franco

[jjasilli]

Don't know how to evaluate this.  Checkout the annexed article.

[jojokeo]

OK, test time... If I connect an 8Ω speaker to my test Hammond transformer, what will be the reflected primary impedance?

Since the same tranny is also used for EL84s (and since there seems to be more AO43 chassis' on ebay?), likely more than 6GW8's  - but in either case my bet is on 8K?!

[jjasilli]

   No betting allowed.  This is math class.      30V on primary > 1V on secondary; turns ration is 30:1.  Impedance ratio is the square of the turns ratio.  30² = 900 Impedance ratio.  Hence an 8 Ohm load on the secondary reflects a 7200 Ohm load on the primary (900 X 8).  OK, fine so you were pretty much right anyway.      But it doesn't count 'cause you didn't do the math.   


A pair of el84's will do, say 15W.  15W = V² / 8 Ohms.  V = √120 = 11V, secondary; 330V primary, presumably, if linear.  Would the OT stay linear in terms of voltage at Max power? Mid power? What about distortion?

[lego4040]

These two videos helped me understand measuring OT’s. I hope they help you out
https://youtu.be/nNlPgirHP48
https://youtu.be/Yzo3A-NywSs

[jjasilli]

These video's are good, but not the info kagliostro is looking for.  He already knows all this.  Kagliostro wants to measure the elements of reactance which is only alluded to (1st video @ 22:49). 


I suspect sluckey's test may not be valid for measuring reactance & distortion, because the primary was left open, or shunted only by the very hi input impedance of a DVM.  I suspect the primary should be shunted by a fixed resistance = to the internal impedance of the intended power tube(s), at each frequency under test.   That way there's current flowing on both windings, with all the reactances in play.  But there's no need to speculate or re-invent the wheel re testing methodology.  That's why recommend hitting the textbooks.  OTOH, maybe reactance testing of an OT is simpler than I think.

[kagliostro]

Thanks to ALL

Unfortunately today I had other things to do and I've a real pain from a molar that broke a few days ago (tomorrow I go to the dentist) so I didn't read or look to any of the link on the latest posts

Excuse me

Franco

[jojokeo]

  No betting allowed.  This is math class.      30V on primary > 1V on secondary; turns ration is 30:1.  Impedance ratio is the square of the turns ratio.  30² = 900 Impedance ratio.  Hence an 8 Ohm load on the secondary reflects a 7200 Ohm load on the primary (900 X 8).  OK, fine so you were pretty much right anyway.      But it doesn't count 'cause you didn't do the math.   

Classic JJ!!! Love it. But if this was math class I wouldn't have to think about things so much - even doing word problems??? Maybe more like Beginning Physics 101?

[PRR]

Good transformers are good from 50Hz to 5KHz. Small voice-only transformers are good 300Hz to 30KHz. (The 30KHz is not needed, it just happens with a small winding and poor bass response.)

The transformer impedance is low at the lower frequency, may be high or low at the upper frequency. So we want to measure in the *middle*. Perhaps 500Hz for big full-bass iron. Perhaps 1KHz for small iron.

The low impedance at low frequency means a load on the source. Slucky knows this and may have picked a strong source (<10r) for his measurements. A 600r signal generator will suck-down very badly on a "8r" winding, and especially in bass.

[jjasilli]

I believe Kagliostro has a decent collection of old unknown trannies, which might not meet modern standards.  Full range OT's arose about 1962.  So I think he wants to accurately measure impedance & imperfections in these creatures: i.e., inductance (which impedes hi's); capacitance (which impedes lo's, except for parasitic capacitance which effectively impedes hi's); & distortion.


My take is that he really wants to "get into the weeds", and is seeking valid testing procedures.

[kagliostro]

I apologize

The dentist had to take off my molar and the thing was not easy and had to proceed by breaking it up and removing it a little at a time

excuse me but given the pain the brain is not very polished to follow the thread at the time

Franco

[sluckey]

Feel better soon my friend.

[jjasilli]

Yes, feel better!  Meanwhile some hitys:


http://www.geofex.com/Article_Folders/xformer_des/xformer.htm


www.aireradio.org/riviste_estere/BVWS/o-p_trasf.pdf

[jojokeo]

My take is that he really wants to "get into the weeds"

This may be a perfect option for him also?! 🤕😷

[lego4040]

Feel better, Ive been there and that pain is like no other. Only when I  had mine done I was prescibed the new pill on the block, Oxycotin.    Slept for three days, got up drove back to dentist and gave it back. I stuck with ibuprofrin after that

[kagliostro]

Today I feel a bit better

I'll go to read all the material you linked

Thanks friends

Franco

[kagliostro]

False alarm, the brain does not work as it should and I can not concentrate the attention,

probably too many pills these days 

Franco

[lego4040]

Lets say I was to use either Uncle Doug’s or El Paso’s method to figure out what OT I have and I dont have a variac. What a good sub? Im sure these transformers draw some current and I have access to stepdown transformers. I think I have 120v to 12 and 120-5v floating around shop cabinets. I know I can use a potentiometer in the circuit to dial in a voltage to make the math easy

[jjasilli]

I think it's easier to do the math than rig up a pot.  Attach OT leads to posts on a breadboard.  Hook up the 5V tranny > OT secondary.  !!!DON'T DUE IT BACKWARDS OR YOU'LL HAVE THOUSANDS OF VOLTS!!! Measure voltages on both sides of OT. Let's say the "5V" side reads 5.8V and the primary side reads 178V.  That's 178 : 5.8. Divide both sides by 5.8.  That's 30.7 : 1.  Bingo! Square both sides for the Impedance Ratio:  942 : 1.  for an 8 ohm load, multiply both sides by 8 = 7535 : 8.

[2deaf]

False alarm, the brain does not work as it should and I can not concentrate the attention,
probably too many pills these days

I have the exact same symptoms and I'm not on any pills.

Of all the inductances involved with output transformers, we are only interested in the inductance of the primary.  The main thing to know about this is that the inductive reactance of the primary winding is in parallel with the primary load that is reflected from the secondary load.  The primary load remains constant over a reasonable frequency range as long as the secondary has a constant load.  The inductive reactance in parallel with the primary load changes radically with frequency.  There are various things that follow from this arrangement.

As long as the inductive reactance is much larger than the primary load, the total primary impedance is essentially the load reflected from the secondary.  As the frequencies become lower, the inductive reactance also becomes lower and at some point it will begin to lower the total primary impedance to a significant degree.  If you lower the impedance ratio in this manner, the output at the secondary will become lower.  When the inductive reactance is equal to the primary load, the frequency response will be down 3dB.  So if you want better bass response, you need a higher primary inductance which usually means more windings and bigger iron.

Inductance is at least one reason for the loss of bass response when you connect a load that has a higher impedance than what is intended to the secondary.  The same inductive reactance will have more effect on the parallel combination when the primary load is increased.   

[jjasilli]

Can you reconcile that with this:  http://www.aikenamps.com/index.php/output-transformers-explained

[2deaf]

Can you reconcile that with this:  http://www.aikenamps.com/index.php/output-transformers-explained

No.  First, I have the attention problem referred to above.  Then, I have no desire to read articles written by experts because it corrupts my pre-conceived notions.  But if you can point to something specific, I will address that. 

[Ed_Chambley]

Can you reconcile that with this:  http://www.aikenamps.com/index.php/output-transformers-explained

No.  First, I have the attention problem referred to above.  Then, I have no desire to read articles written by experts because it corrupts my pre-conceived notions.  But if you can point to something specific, I will address that.

Ok, how can someone increase inductance, reduce leakage while not adding winding capacitance?

[2deaf]

Ok, how can someone increase inductance, reduce leakage while not adding winding capacitance?

I don't believe I commented on this issue, so I have nothing to reconcile with Aiken's article.

[jjasilli]

Here's another article on measuring OT's.  It seems to cite some useful references.

[2deaf]

OK.  I printed out Aiken's article and read the section on primary inductance.  The only differences I see are that he stated an equation and he said "larger core and/or more turns.  I said "more windings and bigger iron".  It is true that more turns on the same core or the same turns on a larger core will produce higher inductance.  I left out the word "or", but I qualified my statement with the word "usually" so that the two statements are compatible.  Also, if you observe amplifiers in the wild, those that have better bass response do tend to have larger iron and more windings.

His formula for the -3dB point does differ from my assertion about inductive reactance being equal to the reflected primary load at -3dB.  However, if the plate impedance is large enough to ignore in parallel with the reflected impedance, then the equation rewrites to f*2*Pi*L = reflected impedance at the -3dB point.  f*2*Pi*L is the inductive reactance at f, so when the inductive reactance equals the reflected impedance, the output will be down 3dB just as I said.

I based my assertion on the assumption (not necessarily a good one) that the equal impedances in parallel will result in half the impedance.  The resulting impedance ratio will result in a voltage ratio that makes the secondary voltage .707 times the original, which is -3dB. 

I am certain that he is right, and I am wrong.  However, both ways results in a similar number and I only mentioned it to illuminate how bass response acts.

   

[2deaf]

Attached are two methods (with some more dubious assumptions) that I used before joining this forum and learning about the voltage test for the primary impedance.  The first method involves the assumption that an isolation transformer will react the same to a reflected primary load as it does to a resistive load.  I set up the test as shown in #1 and adjusted VR-! so that it dropped half the ac voltage.  I left VR-1 at that point and disconnected the unknown transformer.  I then connected another variable resistor as shown in #2.  I adjusted VR-2 until VR-1 was dropping half the voltage again.  I then measured the resistance of VR-2 and called that the primary impedance with a secondary load of 8r.  It yields results that are right around what would be expected on known transformers.

The second method involves the assumption (based on observations) that you can directly read an ac voltage drop across a resistor in series with a coil (of sorts) when the drop is one-half the original.  My general rule is that you cannot directly read voltages when a combination of components involves phase shifting, but this method seems to be an exception because it also yields results that are consistent with known transformers.  When the voltage drop is one-half the original, the resistance of the variable resistor equals the primary impedance. 

[Ed_Chambley]

Ok, how can someone increase inductance, reduce leakage while not adding winding capacitance?

I don't believe I commented on this issue, so I have nothing to reconcile with Aiken's article.

I was not referring to the article either. I have heard some wild things about OTs some true even. Thanks for the input. I lurk a lot around these discussions. I figured you knew more than you were saying.

[jjasilli]

OK.  I printed out Aiken's article and read the section on primary inductance.  The only differences I see are that he stated an equation and he said "larger core and/or more turns.  I said "more windings and bigger iron".  It is true that more turns on the same core or the same turns on a larger core will produce higher inductance.  I left out the word "or", but I qualified my statement with the word "usually" so that the two statements are compatible.  Also, if you observe amplifiers in the wild, those that have better bass response do tend to have larger iron and more windings.

His formula for the -3dB point does differ from my assertion about inductive reactance being equal to the reflected primary load at -3dB.  However, if the plate impedance is large enough to ignore in parallel with the reflected impedance, then the equation rewrites to f*2*Pi*L = reflected impedance at the -3dB point.  f*2*Pi*L is the inductive reactance at f, so when the inductive reactance equals the reflected impedance, the output will be down 3dB just as I said.

I based my assertion on the assumption (not necessarily a good one) that the equal impedances in parallel will result in half the impedance.  The resulting impedance ratio will result in a voltage ratio that makes the secondary voltage .707 times the original, which is -3dB. 

I am certain that he is right, and I am wrong.  However, both ways results in a similar number and I only mentioned it to illuminate how bass response acts.

There's an emphasis in this thread on inductance.  The reactance of any coil, including OT windings, is composed capacitance & inductance along with the static resistance of the coils.  My understanding is, to oversimplify, that induction passes lows but impedes hi's; capacitance passes hi's but impedes lo's; except for parasitic capacitance which also attenuates hi's.  Increasing coil size increases capacitance which also needs to be accounted for; these things are mentioned in the cited articles.  These are the things Kagliostro wants to measure.


It may be that for the relatively narrow bandwidth of el guitar, these parameters may not have much of an effect    (which might be sluckey's point in a related thread).