10K:600 Vs. 15K:600 Ohm Carnhill Transformers

Someone can explain to me the difference between 10K:600 and 15K:600 transformers in a tube preamp output (ECC88)?
I know that with a higher ratio there is less output signal, but there are differences in sound quality?

they should be pretty close,

10K:600 has a  4:1 ratio and 15K has 5:1,

so if the sec has 1000 turns for both,

then the 10K pri will have 4000 turns and the 15K will have 5000 turns,

if wound on the same core and if the interleaving is the same,

then the 10K will have slightly better hi end,

they might have the same bass response even though they are 10K and 15K,

you can find out by checking the primary DC resistance of both,

Cutoff Frequency =  R/L,

R is the DCR of the primary, L is the inductance of the Pri,

we can get an idea for inductance with the 10K and 15K numbers,

if they spec both transformers at say, 20 hz, we have

XL=2 pi F L = 126L, so for XL = 10K and 15K we have

10K=126L, and 15K=126L, so L= 10,000/126 and 15,000/126,

so L=80 Henries for the 10K transformer and 120 Henries for the 15K transformer,

so if you measure the DCR, your cutoff points will be

Cutoff=R/L>  DCR/80 and DCR/120 for the 10K and 15K respectfully,

so if the DCR of the 15K is 1.5 times that of the 10K, cutoff F will be about the same,

if the DCR is equal for both, then the 15K will have better bass,

if you have not bought them yet, then you might find the DCR on the vendors data sheet for the transformers,

Thank you CJ
I have not yet bought the transformers, but in the site (Carnhill/Audio Maintenance) I don't find these specifications

Strawtles said:

Thank you CJ
I have not yet bought the transformers, but in the site (Carnhill/Audio Maintenance) I don't find these specifications

Click to expand...

You won't and to my mind this is the single biggest failing of transformer manufacturers. There are many ways to construct a 10k:600 transformer, some cheap, some expensive, some with good performance some with not so good. Trouble is most transformer manufacturers tell you little more than it's a10k:600 so people naturally ask why shouldn't I buy this cheap one?

Cheers

Ian

yeah it's really hit or miss...

some manufacturers give you the pri/sec inductance, resistance, etc. some don't You'd think it would be the cheaper ones, since they may want to hide some of the less than ideal specs, but edcor even gives pri inductance and ohms.

you could shoot Carnhill an email and see if anything happens,

Sowter is real good about answering emails if you ever have questions on their stuff,

Thank you to all ... I'll try to contact Carnhill

Strawtles said:

Someone can explain to me the difference between 10K:600 and 15K:600 transformers in a tube preamp output (ECC88)?
I know that with a higher ratio there is less output signal, but there are differences in sound quality?

Click to expand...

In terms of affecting the sound, one thing to keep in mind is that the impedance rating on the primary is telling you how it will load the output tube. 

In tube circuits, 10K & 15K should be assumed as values for use with a triode.  For each different tube there is a bias point for a given load at which distortion is at the lowest possible point it can be for that particular tube.  When the loading and/or bias point strays from that the harmonic distortion will begin to increase.  Do note that triodes are very tolerant to these changes and it usually takes a significant change in loading to create grossly unpleasant changes in sound quality. 

As a general rule, triodes like to be loaded with 2 to 5 X their working plate resistance in voltage gain stages but this rule of thumb is seldom observed in output stages where more power is needed.  For a given triode operating with an effective plate impedance of say 8K which is typical,  a 15K primary tends to load a little better for getting lower distortion than a 10K but I wouldn't necessarily always conclude that the 15K will sound better or noticeably cleaner because triodes are fairly forgiving with a wide range of loading tolerance.  You may not even hear it.

I would suggest posting the exact circuit you wish to use the transformers in and let the more advanced members help you calculate the actual changes in distortion you should expect, and whether you will even notice the difference.  There are always a series of trade offs in tube circuits and in this case you will be juggling between changes in harmonic distortion vs loss of gain, which is considered nearly insignificant (only a few dbs).  This is assuming the transformers in question have equal power handling capabilities and don't have vastly out of proportion differences in their primary inductance values.

The question of primary inductance is also a factor for bass response and the 15K may be have a  higher L as CJ explained.  Without specs we sit in the dark.

The question of whether the circuit is direct coupled or capacitor coupled adds different considerations for the inductance values.  Best to see what your circuit is first and go from there.  There are some workarounds for some situations.

he wants to compare transformers for the same circuit.

in other words, keep the generator constant as far as distortion, voltage output (which affects flux density) and source resistance (which will affect matching factor).

you can get more complex by considering core material, flux level, values of distortion in the mid band vs low frequencies, ect,

but for a fixed generator, it still boils down to DCR vs Z of the transformer.

in this case, he has specified the tube type,ECC88,  and he probably wants to hold the operating point constant.

you can change the tube circuit to match the transformer, but we want to hold the circuit constant in order to do an A/B comparison, right?

yes, plate resistance does play a role in calculating distortion, page 209 RDH4,

As CJ said I want only to compare transformers for the same circuit: the EMI REDD 47
I know that the Sowter replacement is a 10K:200 (with a ratio of 7:1) but as I want to use a cheaper transformer (Carnhill) I asked if it is better the 9600:600 (4:1) or the 15k:600 (5:1)
The last one has a higher ratio but a different input impedance than the original one

Strawtles said:

As CJ said I want only to compare transformers for the same circuit: the EMI REDD 47
I know that the Sowter replacement is a 10K:200 (with a ratio of 7:1) but as I want to use a cheaper transformer (Carnhill) I asked if it is better the 9600:600 (4:1) or the 15k:600 (5:1)
The last one has a higher ratio but a different input impedance than the original one

Click to expand...

You asked for an explanation of the difference between a 10K and 15K transformer in tube output circuits . . . . I gave you the best answer I could for that particular question.


That answer applies to your question as best I can tell . . . .  and it applies to the case of use of different transformers for the *same* circuit.   

no, what you said was "wait for PRR"  ;D

which is actually very sound advice,



it's all good,

Kev Ross still likes me,

no wait a minute..... ???

no, what you said was "wait for PRR"  ;D

Click to expand...

To crunch some hardcore numbers, yes. 

I figured he was looking for some explanation of the basics.

I also implied that after the number crunching it may suggest that the differences in audible distortion might not be worth going to the trouble of calculating . . . . but it's always fun to do it at least once.


I coulda said - "hardly any difference in sound"

you coulda said - "the transformer with lower L will have less bass"

But we are trying to be extra helpful to newer members. 

S'all good. 

well, we could analyze 3rd harmonic distortion,

which model?

do we use Rayleigh, Peterson, Welsby, Macfayden or Partridge?

one conclusion is that third harmonic distortion is proportional to the square of the fundamental current and inversely proportional to the square root of the volume of the core,


Peterson say that the third harmonic voltage is proportional to the a, the width of the hysteresis curve, in this solution for the distortion voltage of the transformer, the total source resistance R-G = Rg + Ra, and the load resistance R-L need not be matched,

Macfayden says that  the distortion factor due to a particular material at a given fundamental frequency is a unique function of the ratio R/Zc and the fundamental flux density only, 

also see that the maximum distortion factor is asymptotic to D-max=R/Xp>D,

or we could use V3/V1 =k*aB*Rg/R-L(1+R-G/R-L)^3, where V3 is thrid harmonic voltage and V1 is the fundamental voltage,

Reuben Lee says that distortion due to an elliptical load line (theta=30 degrees) is less than that resulting from the variation of the impedance of the load,

or we could relate total distortion to  phase shift: R/Zc=R/Xp=Dm= tangent theta,


here is a useful table>

the xfmrs in question have dual pri. and secondary;

series pri. parallel sec. for 10k:150, nearly the right ratio

Strawtles said:

As CJ said I want only to compare transformers for the same circuit: the EMI REDD 47
I know that the Sowter replacement is a 10K:200 (with a ratio of 7:1) but as I want to use a cheaper transformer (Carnhill) I asked if it is better the 9600:600 (4:1) or the 15k:600 (5:1)
The last one has a higher ratio but a different input impedance than the original one

Click to expand...

I think a lot depends on the actual load you will apply to the output. When the REDD 47 was designed its output was actually loaded with 200 ohms. That was the EMI standard. Just about everyone else used 600 ohms as standard so a 4:1 transformer would be more appropriate. These days most pro kit has 10K bridging inputs so the question of load is rather moot. I would say use the 4:1 and enjoy the extra gain.

Cheers

Ian

Thank you very much Ian

Strawltes, what kind of E88CC are you going to use?

JJ,  or NOS Tele, Mazda,Phillips...

6DJ8?  7308?  6BQ7a?  6922? 6BZ7?

we need the cathode voltage to do a load line, anybody with a Redd 47 out there? what is the cathode voltage?

Winston O Boogie may or may have not used terminating resistors on the output, original Redd schemo shows 50 to 200 ohms on the output, but i believe the Revolution circuit had the secondary going straight to the output jack,

jeez, Sowter has 30K:600-600 on one page and 10K:200-200 on the other,(click on the Buy 9980 button to read more) and 20 Hz bandwidth in on sentence and 10 hz in another, wtf, over?  i think he means  10K:200-200 and series sec for 30K:600, not 600-600, plus 200 +200 = 800 in transformer math, oh well, maybe he is  letting Drip edit the vintage pages,    if coupling is good it would be 39.2K:800,

i would write the spec as  10K:150//150  or 30K:600 for parallel and series secondaries,



from the size of the can it looks like Brian wound this on a 50 EI  lam using a mix of steel and 50/50 Ni to get the flux level and inductance up there, probably a perm of 14,000 depending on the ratio of Ni lams to steel,0, we can get a rough estimate of pri inductance from the info below,

"9980 Output transformer for REDD47 etc
Recording quality output transformer designed for this excellent amplifier originally used by the Beatles at Abbey Road. The design is based on the original but with our our enhancements. Ratio 7:1+1. 10K/200+200 ohms. Normally used with secondary windings in parallel to give a +18 dBu at 20 Hz but may be series connected for +26 dBu at 20 Hz with secondary windings in series. Very large 50% Mumetal 50% M6 core for high output level, ultra low distortion and exceptional dynamic range. and thick Mumetal can. Colour coded leads. Bandwidth 10 Hz to 50 kH. Can size 45 mm dia x 52.5 mm high mounting grommet (type 9980E) or 2 x M3 holes (type 9980I . Tested with Drip Electronics "Four Seven" PCB.
Schematic
Price: £87.35"

so a DIY xfmr on 50 EI with half steel and half 50/50 on a 1 x 1.5 stack would be 2400 turns pri and 340 turns sec for 250 Henries pri and 5 henries on the 600 ohm sec (150 ohm windings in series) good for 10 hz at a flux density of 14.6 k-gauss @ 26 db, perm of steel and nickel would be about 12,000,

so  120 H at 20 hz way at the top of the page would be 240 H at 10 hz, so we were pretty close,

CJ said:

Strawltes, what kind of E88CC are you going to use?

JJ,  or NOS Tele, Mazda,Phillips...

6DJ8?  7308?  6BQ7a?  6922? 6BZ7?

we need the cathode voltage to do a load line, anybody with a Redd 47 out there? what is the cathode voltage?

Click to expand...

The schematic says the E88CC draws 18mA and the cathode resistor is 200 ohms. The plate resistor is 8.2K.the supply voltage is 290V so the plate sits at about 143V. Load line is from 290V, 0mA to 0V 36mA.

Unfortunately the E88CC curves don't reach 290V but the operating point Ia=9mA, Vg=-3.6V, Vp=143V does look valid. At this point, rp looks to be between 3250 and 3500 ohms which I guess is what you really want to know.

Cheers

Ian

Thank you very much for the detailed informations, but this issue is getting too complicated for my knowledge

I am building my preamps using TESLA E88CC tubes

I have read that other members have built their preamplifier with the EDCOR transformers (10K:600 or 15K:600) but I am considering to buy the Carnhill because I think thay are better than EDCOR
(but cheaper than the Sowter that are too expensive for me)