Help with Transformer Caculator for Impedance Matching

[Ronquest]

I think we could use a Transformer impedance calculator!!!  eh?

Please give feedback and help on what should be included and changed and perhaps when it's finished we can sticky this?

Calling on you guys!  EL34, PRR, Sluckey, Tubenit, pdf64, HotBluePlates, etc....

I've attached an excel file to start, feel free to modify and correct as we go.  I started this in OpenOffice and saved in .xls so formatting may be off.

Thank you in advance for any help with this. 

[sluckey]

Most people can multiply or divide by 2 in their head.

[Ronquest]

Most people can multiply or divide by 2 in their head.

One would think so...  It's not the math that's hard, it's the understanding of what output tap to use with which speaker when the primary impedance is changed.
 You keep answering this type of question over and over and you know it by heart, but others have no clue.  Sometimes I have a clue, but I sit down (again) and draw it out. ( Like: My amp is labeled 8 ohms and my speakers 16 ohms and the transformers 8k, but I'm using it on some EL34s.  Is this OK? )  A calculator/chart would be nice to direct people to.  Maybe I'm just slow, even after reading it 100 times.

[sluckey]

( Like: My amp is labeled 8 ohms and my speakers 16 ohms and the transformers 8k, but I'm using it on some EL34s.  Is this OK? ) 

Short answer... NO

A transformer does not have an impedance. It has an impedance ***RATIO***. In your example that ratio is 8:8000,or 1:1000. If you connect an 8Ω speaker to the secondary that will be transformed to 8K on the primary. If you connect a 1Ω speaker to the secondary that will be transformed to 1K on the primary. A ratio is like an equation. If you do something to one side of the equation you must do the same thing to the other side of the equation. If you do something different on each side of an equation, you no longer have an equation. That's half of the understanding of what an OT does.

The other half is knowing what impedance a pair of tubes wants to be. That comes from tube data and is determined by several operating conditions. But for simplicity, let's just say a pair of EL34s will typically have an impedance of something between 3K and 4K. This is half the impedance of your example. So, if you halve the primary, you must also halve the secondary (otherwise you no longer have an equation). IOW, use a 4Ω speaker to reflect a 4K impedance back to the tubes. 4:4K is the same as 8:8K is the same as 1:1K.

For the tubes to produce their maximum power the transformer must match the tube impedance to the speaker impedance. Any mismatch, either high or low, will cause the power to be less than the maximum.

[pdf64]

…
The other half is knowing what impedance a pair of tubes wants to be.

Just to note that not the right way to be thinking about things. It’s more a case of the load that the valves want to ‘see’ for that given set of anode, screen and control grid voltages (ie their loadline), rather than the impedance of the valves themselves.

That comes from tube data and is determined by several operating conditions. But for simplicity, let's just say a pair of EL34s will typically have an impedance of something between 3K and 4K.
 …
For the tubes to produce their maximum power the transformer must match the tube impedance to the speaker impedance. Any mismatch, either high or low, will cause the power to be less than the maximum.

Maximum AB1 power occurs with the loadline passing through the knee of the Vg1-k=0 anode curve. A pentode’s anode resistance will be several, perhaps many times that loadline impedance.
eg from the standard 250V conditions, EL34 anode resistance is 15k, whereas the suggested load impedance is 2k https://frank.pocnet.net/sheets/019/e/EL34.pdf

The action of the screen grid is to decouple anode current from anode voltage. From that, the anode resistance must be high.

[PRR]

> loadline passing through the knee of the Vg1-k=0 anode curve. A pentode’s anode resistance will be several, perhaps many times that loadline impedance.

Or..... a pentode has TWO "plate resistances". One for low plate voltage and one for high, with your "knee" in between. A happy load tends to be midway between these two resistances.

But for most pentodes that's a wide hill and it really comes down to what voltage and current works for you. Sadly Ohm's Law is not available in the 21st century.

[sluckey]

…
The other half is knowing what impedance a pair of tubes wants to be.

Just to note that not the right way to be thinking about things. It’s more a case of the load that the valves want to ‘see’

To be, to see... Poor choice of words on my part. I was just trying to keep is simple as in multiply/divide by two.

[shooter]

Sadly Ohm's Law is not available in the 21st century.

that's 'cuz it's hidden in 20th century books and we all know books aren't real in the 21st century


this passed through my feed the other day, copied it to my end of world download file


EDIT:
this warning showed up today, keep them together for better results on math tests 

[dwinstonwood]

In an old thread here, PRR used an analogy of a farmer buying a 10 mile power cord to get 10 watts from the power company.
He did all the math to show numerous examples of the power out vs power lost.
I'll try to find the thread. It's maybe ten years old.

[joesatch]

if i have a pair of EL34 and an 8K OT and I would like to deliver 4k to the power tubes  i will connect a speaker half the ohms as the OT output setting?  OT set to 16ohm  into an 8ohm cab?

[sluckey]

if i have a pair of EL34 and an 8K OT and I would like to deliver 4k to the power tubes  i will connect a speaker half the ohms as the OT output setting?  OT set to 16ohm  into an 8ohm cab?

Still at it? Will you ever believe? Oh well, third time is a charm. Or is it, three strikes, you're out!   

[Ronquest]

if i have a pair of EL34 and an 8K OT and I would like to deliver 4k to the power tubes  i will connect a speaker half the ohms as the OT output setting?  OT set to 16ohm  into an 8ohm cab?

Still at it? Will you ever believe? Oh well, third time is a charm. Or is it, three strikes, you're out!   

joesatch, this calculator is for you and for sluckey! 

[joesatch]

if i have a pair of EL34 and an 8K OT and I would like to deliver 4k to the power tubes  i will connect a speaker half the ohms as the OT output setting?  OT set to 16ohm  into an 8ohm cab?

Still at it? Will you ever believe? Oh well, third time is a charm. Or is it, three strikes, you're out!   

halving things sounds logical but i'm not 100% sure if i should be halving the OT output instead of the speaker.

[Ronquest]

In seriousness, what should be included?

Tube to Transformer to Speaker
Turns Ratio
Some general tube specs maybe at a few voltages, ie. EL34 @300V ~ 3400 Ohms, PP and SE

What would be most helpful, without getting too complex?

[shooter]

the OT has NO IMEDENCE
it has a step-up or step down RATIO
in our application, the OT is a step down
so without using any math from the 17 formulas posted earlier;
"fender say; 8k:8 ohm"
"you say 4k:X ohm"
winner with no actual math; 4K is half 8k, therefore X (speaker) = 4ohms

[sluckey]

joesatch, this calculator is for you and for sluckey! 

Not me. I memorized my times two tables in third grade. Only time I get in trouble is when the number has 3 digits.   

[sluckey]

In seriousness, what should be included?

Tube to Transformer to Speaker
Turns Ratio
Some general tube specs maybe at a few voltages, ie. EL34 @300V ~ 3400 Ohms, PP and SE

What would be most helpful, without getting too complex?

Here's a good start.   

[HotBluePlates]

In seriousness, what should be included?
...
What would be most helpful, without getting too complex?

A transformer does not have an impedance. It has an impedance ***RATIO***. In your example that ratio is 8:8000,or 1:1000. If you connect an 8Ω speaker to the secondary that will be transformed to 8K on the primary. If you connect a 1Ω speaker to the secondary that will be transformed to 1K on the primary. ...

In my old[er] age, I'm starting to become grumpy.  I want to tell everyone, "If you have to ask, connect the rated load only."

Sluckey is correct about the transformer levering "impedance" up/down.

It's also true that there is a "desired loading" for a given tube when using a certain power supply voltage, and a certain idle current, and a certain class of operation.

It is true we can change the speaker-load attached to reflect a different-impedance on the primary.  But should we do that?  What's not always considered is the transformer is wound with a certain Primary Inductance such that the "transformer will continue to behave as a transformer, at the lowest frequency of interest."



Huh?

    - Hammond's 1750K is their Super Reverb replacement OT, designed for 4kΩ Primary : 2Ω Secondary.

    - The spec sheet shows the primary Inductance is 12.5H.
    - The guitar's Low E is around 83Hz, and Hammond claims even frequency response down to 70Hz.

    -  The transformer's primary needs to have inductance with a reactance >>4kΩ at the lowest frequency of interest to allow the Turns Ratio to make the primary look like "4kΩ."

    - 2 x π x f x L = 2 x π x 70Hz x 12.5H = ~5.5kΩ  ----> works, but just-works.

    - If we attach "4Ω" to the secondary, we should get "8kΩ Primary" based on the Turns Ratio.  However, the bass will be shaved because the OT lacks the inductance to make this happen.  At best, the OT looks like 5.5kΩ at 70Hz, but perhaps it seems like even less.  Frequency response will taper off on the low end, and we hear "bass-shave."

[Jalmeida]

Here is a OT calculator.

http://www.mh-audio.nl/Calculators/OTLoadResistance.html

Now just get off their lawns so they stop yelling at ya!! 

[Platefire]

I want to point to this article on "Output Transformers" from 1960 by James Moir. I provided a link on my VM 8810 project thread but I don't think anybody noticed it. I think it's very good even though I'm not through it all yet. It's in two parts, so I'm still on part 1. Just like the OP,  this is a part of amp function I also have a hard time wrapping my head around---but I'm getting more understanding, as I go along. The first part of the article talks about the problem getting a voice coil to work hooked directly to the anode circuit of the valve/valves. Makes you appreciate what a great invention the OT was in audio amps. You may not agree but I like it:

https://pearl-hifi.com/06_Lit_Archive/02_PEARL_Arch/Vol_06/Sec_26/1555_Output_Transformers.pdf

[HotBluePlates]

I want to point to this article on "Output Transformers" from 1960 by James Moir. ...

Good find!

The issue I discuss above is Item 3 from "Losses in Transformer" at the end of Part 1, which is explained in depth on the first 3 pages of Part 2.

[Ronquest]

A while ago, I was sorting out a transformer for a build and had a few output transformers from Marshall (Origins 20 & 50) and old Hammonds that where not labeled, so I wanted to get a grip on transformer function beyond just needing the right one.  With a few videos and some web search I set out to demystify some iron.  One transformer after another I realized that figuring out the voltage/turns ratio was straight forward and inductance impedance ratio is not = to turns ratio but, inductance impedance is proportional to the square of turns ratio.  Ad a square and loose 95% of your audience.  I also learned that you can't always trust the label on the transformer or the wire colors....

Transformer choice importance for me.
-watts
-labeled impedance
-orientation
-frequency response

Frequency response for me is generally not high on the priority list, but is by no means unimportant. I tend to like 2nF coupling caps!  HBP has brought this to my attention though, thanks! 

HBP "But should we do that?". Weather we should or not, will not keep us from doing it.  joesatch's question about his 8K with EL34s is a prime example.  Can that EL84(?) transformer and the PT handle the EL34s in his circuit?

So should the calculator include turns ratio, inductance ratio and frequency calculations?  I would think so.
These calculators are readily available, but not all in one spreadsheet.

I realize that these are basic calculations for some of you, but it's nice to have access to some instant numbers that can help give accurate info with the proper input.   I'm grateful for RobRobinette's tube bias calc, rapidtables power calcs, ampbooks, and thesubjectmatter's Power Transformer Current Draw Calculator.  These are tools that help me design and build an amp.  Ohm's law is not lost on me.  I occasionally check the instant answers with my calculator and some scratch paper.  I still look up my formulas for most, as I don't always remember them on next new build months later.  I can build a Fender, Marshall or Vox style amp without a schematic or design one from scratch on paper and have it work without a single calculation, as many of you can too. I had 2 years of EE some time ago and I'm far from any recollection of my calculus classes.  My career path took a different direction and the EE was pretty much sidelined, but the engineering part of me still exists, just ask my wife.   She's says I can't just do something without redesigning the whole thing.

There's a wide range of amp builders and individuals on this forum, ranging from strict engineer to the weekend modder, the engineer wants all the calculations with consideration of every component and the new modder is only seeking the mod, at least till his amp smokes.  Some newbies get the bug, dive in and read posts from the engineer group while hanging on your every word.  I've read post here from some of you nearly 20 years ago, did you know as much then as now?

Thank You to all the guys who help to keep us straight, your time and effort is most appreciated!

[sluckey]

inductance ratio is not = to turns ratio but, inductance is proportional to the square of turns ratio.

You keep saying inductance but I think you really mean impedance.

I've read post here from some of you nearly 20 years ago, did you know as much then as now?

I probably knew more about electronics 20 years ago because I was still actively engaged in my career. I've been retired for 12 years now and I have probably lost a lot. Can't say how much. You can't know what you forgot. And now that amp building has faded away and been replaced with wood turning, I'm sure I'll forget even more electronics stuff as the noodles absorb woodworking stuff. Only got so many wrinkles.   

[Ronquest]

You keep saying inductance but I think you really mean impedance.

The internet is full of lies.  I think I mean impedance but the web tells me it's inductance.  LOL   I'll change my post above, I don't want to be part of the lie.

[PRR]

...You keep saying inductance but I think you really mean impedance....

Inductance times FREQuency IS impedance.

In a small world which mostly stops at 82Hz, it may be a distinction without a difference.

The answer is, as you say, probably simple to figure. Phrasing the question seems to be the hard part.

[shooter]

Can't say how much.

Back then I knew amazing things, since they have little to no relevance to my current life, I set them free!!!
every now and again HBP or PRR will make me go look under the bed, but the wow factor is gone.


my new fascination is the speed at which the 20th century became irrelevant, relegated to dinosaur bone heaps