VanGenz

G7 - almost finished - which EF86?
« on: August 21, 2012, 03:57:46 PM »
After several years of builing I am almost finished with my G7/Equinox Michrophone. The body is Equinox EQU47 (U47 copy) and the capsule is Equinox Q47-D. I have changed some resistors to lower the voltages to the capsule in the G7 circuit. Had problems getting the +B voltage up to 160 volt, changed the second trafo from a 220:15 to a 220:12 volt still the +B voltage (loaded) is "only" +145volt. Had problems with massive hum in the first test. But when I connected the "Heather 0" to the common inside the mic the hum was gone. In the drawing of the mic-pcb the common and the Heather-0 is connected, but they were not on the pcb. Have borrowed a Wunder U47 clone and compared the two michrophones. When talking into the mics I can hear no difference in the sound, so I guess the +145 volt is ok for me. But the noiselevel (steady noise - no hum)is 15-20 db worse in the G7. The tube is a Electro-harmonix (made in russia) EF86EH. I understand I could reduce the noise with a better tube. There are "loads" of different EF86 tubes on Ebay. Which tube should I go for to get the best signal-noise performance?


gyraf

Re: G7 - almost finished - which EF86?
« Reply #1 on: August 22, 2012, 01:03:01 AM »
145V is all you need - actually you could (and maybe should) go as low as 120V for some capsule types (like CT12).

As for tube noise at very-high input resistors, this is an unpredictable parameter in current (new) tube manufacturing - you may need to go through a few to find a quiet one.

What I have found is that the PF86 (nos, same tube, different heater voltage, much lower price) in general gives better performance - specially for the price. You need to change heater voltage in the psu to use this.

Jakob E.
..note to self: don't let Harman run your company..

ruffrecords

Re: G7 - almost finished - which EF86?
« Reply #2 on: August 22, 2012, 06:16:10 AM »
There is no one version of the EF86 that is guaranteed to give lowest noise. Even when you pick a reputedly low noise type, there is no guarantee just how low the noise will be. For this type of application you need to buy quite a few and select the one with lowest noise. Expensive but that's the way it has always been done.

Cheers

Ian
www.customtubeconsoles.com
https://mark3vtm.blogspot.co.uk/
www.eztubemixer.blogspot.co.uk


'The only people not making mistakes are the people doing nothing'

VanGenz

Re: G7 - almost finished - which EF86?
« Reply #3 on: August 22, 2012, 01:49:14 PM »
Called a friend of my who has serious GAS...(gear acguisition syndrome) and he came over with 7 tubes...3 x Ediswan EF86, 2 x Philips mimiwatt and 2 x Mullard PF86. Tested each of the EF 86 first, with 6.3 volt glowing. One of the Philips tubes were dead (the two philps tubes were used, the other 5 were nos) Tested for 5 minutes each and the noise level was the same. Adjusted the heather down to 4,5 volt and tested both of the PF86. Thes same noise-level, Will leave each of the working tubes on for 24-48 hours to check which "burns in" the best. Noticed that the +b voltage fell from +145 to +135 without the sound changing.

Re: G7 - almost finished - which EF86?
« Reply #4 on: August 22, 2012, 05:39:28 PM »
15-20 db more noise than an other tube mic sounds huge.  And when noise level stays same regardless of the tube it makes one think that there's something else going on than a bad tube. How is the output level?

dmp

Re: G7 - almost finished - which EF86?
« Reply #5 on: August 22, 2012, 05:49:49 PM »
Have you cleaned the board carefully to remove any solder flux from the high impedance areas (near the capsule)?
This can cause hiss.

VanGenz

Re: G7 - almost finished - which EF86?
« Reply #6 on: August 22, 2012, 06:47:52 PM »
I am also suspecting that the tube is not the problem here... Have cleaned the mic-pcb quite good but will clean it bether tomorrow. Regarding output I have to give 14-15 db more gain on the G7 channel than on the Wunder channel to get the same output level. I guess this a symptom that something is not correct in the mic-circuit. If I set equal gain on both mics the noise floor is just 2-3 db higher on the G7, but the output is of course lower.  But when I speak into both mic's I can really not hear any diffreence in sound. I have a Audio Technica 4047/SC and I can tell the difference between this and the Wunder.

baadc0de

Re: G7 - almost finished - which EF86?
« Reply #7 on: August 23, 2012, 03:42:30 AM »
I liked Telefunken EF806S best. They live in my preamps. The Telefunken EF86 are noisier, but sound great too. I have tested Siemens, Tesla, JJ and a russian equivalent. The russian one has good noise levels, but I didn't like the sound.

VanGenz

Re: G7 - almost finished - which EF86?
« Reply #8 on: August 23, 2012, 06:43:55 AM »
Could anyone tell if there is a way to see what is front and back on the Equinox Q47-D capsule?. I see no color or marking difference. I find no information from Equinox. Does it matter? Cleaned the mic-pcb more and changed the 2,2uF capasitor from 100volt to 250 volt. But the noise and outputlevel is the same. Should I expect same outpulevel from the G7 as other tube-mics?

Re: G7 - almost finished - which EF86?
« Reply #9 on: August 23, 2012, 05:31:58 PM »
It doesn't matter. I believe you should expect the output level to be in the same ballpark than the Wunder (an EF14 plate fed circuit I believe). I'm not wise enough to help you further, but I would double check everything first and after that trouble shoot systematically.


VanGenz

Re: G7 - almost finished - which EF86?
« Reply #10 on: September 05, 2012, 04:49:15 PM »
Found two faults and is very close now! The reason the signal was low was because I had soldered one of the balanced outputs (should have been pin 3 two the chassis ground in the XLR-3 pin out. Also found a bad contact/solder in one resistor. I am now checking several of the EF/PF86 tubes for the lowest noise. There is still more noise, and especially in the low bass (rumble). When tapping on the body of the Wunder and the G7/Equinoxe the G7 is much worse and give an extreme sub-bass-noise. Some of the same difference if I make a "Ph" close to the mikes. Could I reduce this rumble-noise/sub-bass respons by reducing the "1n Polysteren" to 470 or 330 pF?

useme2305

Re: G7 - almost finished - which EF86?
« Reply #11 on: September 05, 2012, 05:05:30 PM »
the higher noise could be the result of non proper shielding. make sure the body and the mesh are grounded. also, different mesh materials have different conductivity which lead to different shielding features even when they are grounded properly. different mesh structures also behave slightly differently i believe. do some tests!

dmp

Re: G7 - almost finished - which EF86?
« Reply #12 on: September 05, 2012, 05:08:49 PM »
Quote
Could I reduce this rumble-noise/sub-bass respons by reducing the "1n Polysteren" to 470 or 330 pF?
You shouldn't have to change any component values to get the mic working without issues.
Do you get the same noise with different tubes? That is likely since the rumble is influenced by tapping the mic. Or else you have a bad joint / ground.

gyraf

Re: G7 - almost finished - which EF86?
« Reply #13 on: September 06, 2012, 12:55:21 AM »
the very-low frequency response in the G7 is on purpose - not low-cutting the circuit gives subjectively much better lo-freq impulse response / coherence (and I like that).

That said, you shouldn't use the G7 for vocals without a pop filter.

If you want to introduce a low cut, work with the 1G resistors. Cutoff frequency will be capsule capacitance into the two 1G's parallel'ed. Lower values (say 100M) will give higher lowcut frequency.

Jakob E.
..note to self: don't let Harman run your company..

Re: G7 - almost finished - which EF86?
« Reply #14 on: September 06, 2012, 06:00:29 AM »
If you want to introduce a low cut, work with the 1G resistors. Cutoff frequency will be capsule capacitance into the two 1G's parallel'ed. Lower values (say 100M) will give higher lowcut frequency.
Don't do that.  Going from 1G to 500M or even 100M will reduce the bass but also increase hiss.

This may be swamped by the noise of the EF86 but if it works OK with 1G (Anode voltage is correct), it is better to reduce LF elsewhere.  Maybe the cap to the transformer.

Your Equinox capsule might simply have better bass than the Wunder.  Are both mikes on identical shock mounts?

gyraf

Re: G7 - almost finished - which EF86?
« Reply #15 on: September 06, 2012, 11:13:10 AM »
Actually, I've had cases of significantly reduced noise from some specimens of EF86's when experimenting with different (lower) capsule loads.

It seems that at least some of the EF86's really like a lower grid resistance.

Jakob E.
..note to self: don't let Harman run your company..

dmp

Re: G7 - almost finished - which EF86?
« Reply #16 on: September 06, 2012, 11:42:15 AM »
Quote
Going from 1G to 500M or even 100M will reduce the bass but also increase hiss.
I thought the higher the resistance, the higher the hiss. Resistor thermal noise scales with temperature and resistance.
Lowering the resistance then would reduce the bass, but also reduce the hiss. Or does the grid leakage current in the tube generate noise, ie shot noise?

I interpreted the OP as having a problem in his build. I have a G7 that has good bass response but does not have a problem like this:
Quote
more noise, and especially in the low bass (rumble)
. And he possibly has a microphonic problem:
Quote
When tapping on the body of the Wunder and the G7/Equinoxe the G7 is much worse and give an extreme sub-bass-noise.


Re: G7 - almost finished - which EF86?
« Reply #17 on: September 06, 2012, 08:42:01 PM »
Quote
Going from 1G to 500M or even 100M will reduce the bass but also increase hiss.
I thought the higher the resistance, the higher the hiss. Resistor thermal noise scales with temperature and resistance.
Lowering the resistance then would reduce the bass, but also reduce the hiss. Or does the grid leakage current in the tube generate noise, ie shot noise?
The capsule capacitance forms a LP filter for the noise.  The noise is increased at subsonic frequencies but reduced at audio frequencies.  The total noise power (surprise surprise) remains the same.

So the Grid/Gate resistor is always determined by noise rather than LF response considerations.

The definitive treatment of this subject is Great Guru Baxandall's 1968 WW article.  http://tech.groups.yahoo.com/group/micbuilders/files/Ricardo/
You have to join.

The limit with valves is Grid Current.  If the resistor is too large, the Grid Current will change the voltage on the Grid and the operating conditions on the valve.  This will vary from valve to valve and maybe over time too.  But if the anode voltage is correct, the operating conditions are probably OK so you can keep the 1G.

Jakob, I think this falls down if the current noise of the valve is high.  But those would not be good valves to use in a high Z LN circuit.

I need to say I am only a pseudo valve guru and may be pontificating from the wrong orifice.  But I once made a living designing FET condensor mikes.

VanGenz

Re: G7 - almost finished - which EF86?
« Reply #18 on: September 07, 2012, 07:07:19 PM »
Thank you for all the response! Have been busy (playing guitar)so thats the reason for late answer. I have also finished two Equinox bodies/capsules with the MK47-kit and these behave fifferent. One has noisy tubes, and I have ordered 10 x 6028 nos tubes to fix this. The other is working ok and is nearly as quiet as the Wunder mik. When tapping the body ofthe  two MK-47 mic's they are not as sensitive as the G7. But the noiselevel is still higher in the G7. There is almost no hum in the mic's (too much noise to really hear this in the G7) and there is infact a little more hum in the Wunder than the MK47. I ended making the mic-cable with shield and earth soldered to a common pin in both ends of the tube-mic-cables (insted of letting the shield "hang" in one end as I first thought). "Are both mikes on identical shock mounts?" Yes I test with two equal. But I guess the PF86 and EF86 tubes I have tested so far is more microphonic than the EF800 (propably) sitting in the Wunder. Easier to test more now when I have 4 good tube-mics to compare.

VanGenz

Re: G7 - almost finished - which EF86?
« Reply #19 on: September 10, 2012, 11:27:10 AM »
I used 470Kohm and 282Kohm (instead of 2 x 470Kohm) as suggested from equinox to get 60 volt polarization of the capsule in the G7 power.. This only gave 47 volt so i used 420Kohm instead og 282Kohm. This gave nearly 60 Volt (59,5) and the mic became a little less "michrophonic" when touching the body. Still more noise than the Wunder and the MK47. But when I change the pattern to "eight" the noise becomes almost as low as the MK47. I suspected the pattern-switch (which was an old an oxyded Lorlin) changed this to a new, but still more noise in omni and cardoid than in eight. The level in eight and cardoid seem equal (lower in omni which I guess is normal) The noiselevel in the Wunder seem equal in all three patterns. I have also changed the two resistors in the G7 power from 2 x 100K to 40K and 2 x 50K. Since the +160volt in reallity is +137volt the 40/50/50K divider does not provide the theoretically correct eight and cardoid pattern-voltage, but I do not understand why the figure-8 voltage (around 98 volt) gives lower noise than the cadoid (around 49v) and omni (0 volt).....?


 

Related Topics

  Subject / Started by Replies Last post
4 Replies
3029 Views
Last post July 16, 2004, 05:20:56 AM
by gyraf
9 Replies
3485 Views
Last post July 23, 2004, 05:45:22 PM
by nacho459
6 Replies
2426 Views
Last post April 07, 2006, 12:16:50 AM
by pxpx83
31 Replies
6505 Views
Last post September 03, 2006, 06:12:40 AM
by Rob Flinn