Eros with Flux magnetics ER head

[ALEXZ]

Hello,
I've got my RS-1500 headblock with Flux head installed by JRF (thank-you-John!).  Measurements with the TP test tape shows that 1khz and 10khz signals have identical amplitude,  16 khz is about 3 dB up and 30 hz is about  5 dB up.  All this with IEC IQ. Should I try to adjust 35ms to make high end flat and what to do (if any) with the 30hz ? 
Thank you,
Alex

[Doc B.]

Try installing a 100K ohm trim pot wired as a variable resistor (center pin connected to one end pin) in place of each 47K input grid resistor (T19 to T20, T26 and T27). This should allow you to adjust the head damping and level off the top end response.

I'm going to let PJ weigh in on treatment of the low end rise.

[Grainger49]

As much as that curve would please my ears I understand your desire for flat response.

You want a ten turn trip pot, if that wasn't obvious.  I like the big ones that are easy to tweak but they have easily measurable inductance.  Most have the inductance printed on the side.

If you don't need higher than 47k then a 50k pot will give you much higher, double, the resolution.

[ALEXZ]

Thanks Dan,
Currently I have 100K (actually 91k, that was adjusted for the stock head) input R. Interestingly, the FM head is the same 200mH as a stock one, I was under impression that there are no reasons to change input R, now I

[Doc B.]

I don't think you need a ten turn pot, it's not that precise an adjustment. I'd just rig a small 100K trimpot and see if you can find a setting that levels the HF. Since it's rising only above 10kHz my hunch is you don't want to make any changes at the 35uS pole (4500Hz).

On the bottom end the IEC "standard" is a theoretical steady slope to infinity at 0Hz (!). So it's basically left up to the designer of the EQ where he will create a -3dB point for the bottom end so you don't need infinite gain (!). I don't recall exactly what PJ chose for that point, but I'm pretty sure it's below 30Hz, and it might be possible to tweak a value somewhere to bring that roll off up a bit to compensate for the rise.

[ALEXZ]

Ok.  The story ended with 49k / 55k  input Rs.  Top end is not perfectly flat,  10k is a little bit below base level (1k) , 15k is  at base level, and 16 a bit up (it's all in a level of  less then 1dB ).
For those curios were is 16k signal on the TP tape: Rs1500 has otherwise useless pitch control plus/minus 5%. In my case it's bit more and full speed ahead brings 15k to 16k
Thank you, Dan for you help, works like a charm.

[Paul Joppa]

The LF response is unexpected. 5dB is a lot for the head bump, especially since the FM head is said to be extra wide and shallow to minimize such a thing. The Eros rolls off the boost at 15Hz, so at 30Hz it should be 1dB DOWN relative to the IEC curve.

You can implement the NAB rolloff at 50Hz - I'll have a look at the circuit again to be sure how to do that, and I follow up if you decide to go that way. That will pull the 30Hz down  about right, though there may be a small droop at 60Hz. But this seems odd enough that if it were mine, I'd want a bunch of answers to other questions first. What - exactly - is the test tape? i.e. do we really know that it does not have the 50Hz NAB boost? Is it full track or half track (i.e. is there a fringing effect)? What - exactly - is the meter used? i.e. are we sure it is accurate below 60Hz? There are many ways to get erroneous measurements at the frequency extremes, and I'd like to be sure we are addressing a performance problem rather than a measurement problem.

[Doc B.]

If the tape is our test tape it is 2 track and measures flat at 30Hz on the majority of the machines I have seen with Flux ER heads.  That is to say, this sounds like a known test tape, head and preamp configuration, that usually measures flat at 30 Hz. The test tape measures flat on my Nagra T which has similar inductance, though it is not a Flux head. I'm glad you said you need to look at the map a bit to determine what value to change. It's been a while since I've dealt with it, and the somewhat trick passive EQ we use had me scratching my head a little.

I will check with Paul S to see if he has noticed any deviation at 30Hz on any of our test tapes. As far as I know this has not been an issue.

[ALEXZ]

I guess I

[Doc B.]

The 0 dB at 15K and +2 at 16K on the FR head seems odd. I would think you might see more like .5dB max difference between frequencies that close together. Are you using a scope to take the measurements? Sound card/computer? The big upward trend at each end of the frequency range makes me wonder if there is something in the measurement setup that is influencing the results. Not that I have a clue of what might cause it, as we usually find problems like high capacitance test leads rolling off the frequency extremes rather than boosting them. Is the 1/4 track head stock? If so, we have certainly seen worn heads that measure rolled off at 10K and above like that.

Regarding the aggressive sound - was the old 2 track head rolled off on top? If so, it might not be the electronics, but rather the transport that is making it sound a little bright and aggressive now that you are hearing the full frequency response on top. Scrape flutter can add some small bit of harshness to the sound. The tape path mod we came up with helps a lot to reduce that effect if you haven't done it yet.

The only other possibility that comes to mind is that it might be the tubes that sound a little bright now that you have the extended top end response. Both the 6922 types and the EF86 types come in a lot of different sonic flavors. As much as we try to level the playing field with the circuit design, the unfortunate truth is that expensive, highly sought after tubes usually do tend to sound the smoothest and most resolving.

[ALEXZ]

Hi Dan,  thank you for the response.  
Here is the environment:
I

[Doc B.]

OK, sounds like good tubes! Not familiar with the sound of the EF804S, but assuming it is close to the EF806s it should not be the problem. 20% doesn't seem that bad, that's less than 1dB difference if I understand correctly. Could be there is still some resonance above 16kHz and that the head damping trim just doesn't quite flatten it out. Maybe a different head cable would help.

Aligning 1/4 track heads is tough. Tone tapes are usually either full track mono or (like ours) 2 track stereo. So it's difficult to get the skinny little bands on the 1/4 track head well aligned with them. We usually do a rough alignment with the tone tape to get azimuth as good as possible and then listen to a known 1/4 track tape and tweak the height adjustment for the best HF response. That way you know the bands are lined up with the tape tracks as well as possible.

[Grainger49]

Dan,

Alex pointed out the EF805S as a replacement for the EF86 that requires a socket rewire, i.e. same spec different pin assignments.  It is listed as such on the Tube World site.

[ALEXZ]

Dan,
- EF804S is a twin brother of EF806S. 
- 20% in amplitude is ~ 1.6 dB

[Paul Joppa]

Just a word or two of caution about getting carried away with analysis in magnetic devices. We speak of inductance as if it were a constant, but it's not. In something with a core, like a tape head or a phono cartridge or a transformer, the inductance varies with the permeability of the core material, and that permeability is a function of both magnitude and frequency of the exciting field, as well as the current state of magnetization (e.g. intermodulation of low and high frequencies). The effective inductance at high frequencies will be smaller than the low-frequency-rated 200mH, and the resonance frequency itself will shift with signal level a bit.

The terminals also show a series low resistance (the wire resistance) and a parallel high resistance (eddy currents in the core material) - both of which vary with frequency due to the skin effect. And there are hysteresis losses, which appear as if they were a resistance but are non-linear and, like permeability, a function of magnitude, frequency, and magnetic state. These loss mechanisms damp the resonance, along with the preamp input resistance.

And of course the coil has self-capacitance plus capacitance of the wires to the core and shell; it's not just cable capacitance making the resonance happen.

Well, those are the confounding effects off the top of my head; I'm pretty sure I've forgotten a few and I have faith that more will be discovered eventually.