Telefunken M15 frequency response.

[sigitask]

Hello! Im trying to allign M15 record-playback frequency response to maximum flatness as is possible. Tape SM468, bias ower 3dB. Have drop on 30Hz, what I dont like. Is it normal for M15 or it is some malfunction?

https://www.flickr.com/gp/143324731@N02/Vhvb3P

[kaimex]

Have drop on 30Hz, what I dont like. Is it normal for M15 or it is some malfunction?

Could you be more specific ?
How about a preliminary frequency response plot ?
At the low end every system has some drop.
Tape recorders even have something special called "head bumps" or "contour effect", resulting in an "undulating response" (McKnight).

Regards Kai

[t20]

RTM recommends 4 or 5 dB overbias depending on the tape speed: https://www.recordingthemasters.com/wp-c...COM-15.pdf .

Did you plot the frequency response of the reproducer part of the machine using a calibration tape?

Regards,

Nelson

[kaimex]

Hi Sigitas,

beg your pardon that I overlooked the flickr link.

That frequency response looks quite good to me. -3 dB at 30 Hz isn't bad.
Peter will know better what is possible with a M15 at 15" with SM468.

What do you expect ?

Regards Kai

[kaimex]

Hi Sigitas,

in the meantime I have looked at the schematic of the M15A BC WV1 playback amp.
Did you already try to get the 30 Hz level up by adjusting the "Tiefen 38" trimmer R27 ?
By setting it to the maximum value (1Meg Ohm) you will get higher 30 Hz amplitude.
You could even get more by additionally replacing R23 with a higher valued resistor.
It looks as if the flat portion of the frequency response can be even stretched down to 23 Hz (-3 dB near 17 Hz) by replacing capacitor C2 (1000 µF) by a smaller value (700...600 µF). These capacitors, however, usually have high tolerances and the proper value might also depend on the open loop gain of the operational amplifier. But if you want to push low frequency response to the limits, this might be an area for experimentation.

Regards Kai

[Peter Ruhrberg]

in the meantime I have looked at the schematic of the M15A BC WV1 playback amp.

Sigitask's machine is an M15 with V396/397 amplifiers.
EQ section of V397:
   

Peter

[kaimex]

Hallo Peter,

thank you for the correction.
What a pity, this amp will behave differently.
I'll check its frequency response in simulation as soon as I find time...

Regards Kai

[Peter Ruhrberg]

RTM recommends 4 or 5 dB overbias depending on the tape speed:

Overbias values also depend on the gap width of the recording head, as shown in the Telefunken manual:

   

Best, Peter

[kaimex]

Good news:

The V397 amp behaves similarly:
Increasing the "Tiefen 38" trimmer R17 to its maximum value (1 MegOhm) and lowering C4 from 250 µF down to 125 µF will result in an Omega-correction close to the theoretical 1/f response down to ~24 Hz.
Otherwise (e.g. with R17 set to the mid value 500k) you will get -3 dB @ 30 Hz.

Regards Kai

[sigitask]

Overbiasing depends on tape type and record head gap. There is not ferrite heads, so it is probably vacodur 18um. So ichose 3db.

[Peter Ruhrberg]

There is not ferrite heads, so it is probably vacodur 18um.

Do you happen to have head types available? They should be printed on the shield:

AB 21 & WC 30-3
   

[sigitask]

Hi Sigitas,

in the meantime I have looked at the schematic of the M15A BC WV1 playback amp.
Did you already try to get the 30 Hz level up by adjusting the "Tiefen 38" trimmer R27 ?
By setting it to the maximum value (1Meg Ohm) you will get higher 30 Hz amplitude.
You could even get more by additionally replacing R23 with a higher valued resistor.
It looks as if the flat portion of the frequency response can be even stretched down to 23 Hz (-3 dB near 17 Hz) by replacing capacitor C2 (1000 µF) by a smaller value (700...600 µF). These capacitors, however, usually have high tolerances and the proper value might also depend on the open loop gain of the operational amplifier. But if you want to push low frequency response to the limits, this might be an area for experimentation.

Regards Kai,

Dear Kai, Yes I made LF adjustment with tiefe potentiometer, but I can't elevate frequencies below 40Hz, I can elevete from 80 and lover, and this elevation makes bump on 60-70 hz about 3 db and its to much, big defference between 60Hz and 30 Hz in any case. The same more les about 4-5 db difference. Changing components can be solution, but art the moment I like to understand how it was with standard components version.

RTM recommends 4 or 5 dB overbias depending on the tape speed: https://www.recordingthemasters.com/wp-c...COM-15.pdf .

Did you plot the frequency response of the reproducer part of the machine using a calibration tape?

Regards,

Nelson

Dear Nelson, I put 3 db overbias, because RTM recomendations is on Studer heds, what gap is twice smaller as M15 vacodur heads, so I made 3 db. Playbach channel was set with MRL TEST TAPE on 100, 1k, 10k, and 16 k Hz, +- 0,5db.

[sigitask]

Dear Kai,
C4 works for NFB, correct? So to decrease it is intersting solution, will try it, thank you.

[kaimex]

Hi Sigitas,

what is "NFB" ? I am not familiar with that abbreviation.
C4 sets the frequency at which the gain of the playback equalization begins to drop towards lower frequencies after the 1/f slope of the Omega-correction. It also has an effect on the "sharpness" of the transition. With the present value (250 µF) this frequency is considerably below 20 Hz and the transistion is very gradual. A lower value raises this frequency, makes the transition sharper and the slopes above (and below) the maximum more linear. In the simulation 125 µF resulted in maximum equalizer gain around 20 Hz and a slope very close to 1/f above 24 Hz, which should allow to get a flat frequency response without an upward bump as you observed when trying to get the 30 Hz level up. (Even lower values of C4 will lead to a frequency response with a sharp resonance which is undesirable. So you have to check, which value leads to the best compromise.)
This assumption, however, expects that there are no other highpass effects in the audio system with lower corner frequency near 30 Hz.
To get a clearer picture what the playback preamp does for equalization, you might want to checks its frequency response ditrectly without tape, by feeding in a test signal e.g. at the transformer input. The low frequency repsonse will depend on whether you feed in current (high source resistance) or voltage (source resistance very low).

Regards Kai
PS.: While brushing my teeth after the evening-snack, it suddenly flashed through my brain: NFB = negative feedback. How couldn't I know right in the first place ? Sure, every"thing" down there at the emitter of Ts1 is negative feedback.
PS2: You can avoid the complication of coupling through the transformer by coupling the test signal with a series resistor and a DC-separation capacitor to the emitter of Ts1. The resistor value should be chosen as high as possible (xx kOhms) with the maximum output level of your signal generator, the capacitor should be chosen such that the high-pass corner frequency of the R-C is somewhere around 3 Hz or less in order not to change the observed equalizer response above 20 Hz.

[sigitask]

Hi Sigitas,

what is "NFB" ? I am not familiar with that abbreviation.
C4 sets the frequency at which the gain of the playback equalization begins to drop towards lower frequencies after the 1/f slope of the Omega-correction. It also has an effect on the "sharpness" of the transition. With the present value (250 µF) this frequency is considerably below 20 Hz and the transistion is very gradual. A lower value raises this frequency, makes the transition sharper and the slopes above (and below) the maximum more linear. In the simulation 125 µF resulted in maximum equalizer gain around 20 Hz and a slope very close to 1/f above 24 Hz, which should allow to get a flat frequency response without an upward bump as you observed when trying to get the 30 Hz level up. (Even lower values of C4 will lead to a frequency response with a sharp resonance which is undesirable. So you have to check, which value leads to the best compromise.)
This assumption, however, expects that there are no other highpass effects in the audio system with lower corner frequency near 30 Hz.
To get a clearer picture what the playback preamp does for equalization, you might want to checks its frequency response ditrectly without tape, by feeding in a test signal e.g. at the transformer input. The low frequency repsonse will depend on whether you feed in current (high source resistance) or voltage (source resistance very low).

Regards Kai
PS.: While brushing my teeth after the evening-snack, it suddenly flashed through my brain: NFB = negative feedback. How couldn't I know right in the first place ? Sure, every"thing" down there at the emitter of Ts1 is negative feedback.
PS2: You can avoid the complication of coupling through the transformer by coupling the test signal with a series resistor and a DC-separation capacitor to the emitter of Ts1. The resistor value should be chosen as high as possible (xx kOhms) with the maximum output level of your signal generator, the capacitor should be chosen such that the high-pass corner frequency of the R-C is somewhere around 3 Hz or less in order not to change the observed equalizer response above 20 Hz.

Hello Kai,
yes NFB is negative feed back So I will try to play with C4. Thank you for simulation and detailed explanation. Im little bit worry that head input transformer can give this problem. Also Im thinking can this problem start in record amplifier, or maybe heads can be cause.

[sigitask]

There is not ferrite heads, so it is probably vacodur 18um.

Do you happen to have head types available? They should be printed on the shield:

AB 21 & WC 30-3
[attachment=34507]

Hello Peter,
M15 heads is WC29 and AC29, is it vacodur?
I dont understand how to attach picture here?

[sigitask]

There is not ferrite heads, so it is probably vacodur 18um.

Do you happen to have head types available? They should be printed on the shield:

AB 21 & WC 30-3
[attachment=34507]

Hello Peter,
M15 heads is WC29 and AC29, is it vacodur?
I dont understand how to attach picture here

[Peter Ruhrberg]

M15 heads is WC29 and AC29, is it vacodur?

Yes.
According to specifications, their mechanical gap widths are 20 (AC) & 6 µm (WC).

Peter

[kaimex]

Im little bit worry that head input transformer can give this problem. Also Im thinking can this problem start in record amplifier, or maybe heads can be cause.

What do the specifications of the M15 promise ?
The successor M15A ist specified with 30 Hz - 16 kHz +- 1.5 dB at 15"/s.
The frequency response you have shown only had 1 dB less at 30 Hz and had ~35 Hz to 65 Hz within 3 dBs and the "rest" up to more than 20 kHz also within +1/-2 dBs.
This isn't something, that one would normally call a problem.

Regards Kai

[Peter Ruhrberg]

What do the specifications of the M15 promise ?

Same as M15A:

   

Peter

[kaimex]

Hi Sigitas,

the frequency response of the playback pre-amplifier (without transformer) has at least one octave lower - 3dB "cut-off" frequency than found in your measurement, if all parts have the values documented in the schematic. The transformer will add some 1st order high-pass effect depending on the sum of the series resistances of playback head and transformer primary and their inductances. In case the capacitance of C4 (250 µF) has dropped down to ~45 µF, a cutoff frequency near 30 Hz would be observed. If non of these potential causes for the discrepancy can be confirmed, there must be some other 2nd order ~30 Hz high-pass in the system , e.g. in the recording channel(s).

regards Kai