Mic Fet booster with stepped HPF and line drive

[Tubetec]

Here we see the proximity effect of a 58 mic ,
The simple HPF gives near perfect compensation in the low midrange , while lower frequencies are filtered off a bit more sharply .
Interesting that the HPF curve starts to flatten out at more extreme cut settings , this leaves the sound lacking punch ,sounding unaturally weak , where the lower rounded curves have a much more natural balance .

Its useful being able add or vary the load resistance externally to fine tune the filter frequency ,
Ill try a running another set of curves with 25kohms load later ,



It brings us nicely on to the buffer/line driver ,
As a starting point the single rail transformer coupled class b transistor amp shown below looks good ,
but using an ungapped API style output transformer and biased class A ,
preferably able to work over a wide supply voltage range ,depending on the output drive requirements .

Can the output transformer be wired into the emitter circuit below?
Is it the same as the cathode follower equivalent of a transistor ?
does it have no gain(because of local feedback) ,but low output impedence and low distortion ?

classbamp.JPG (25.48 KiB) Viewed 1904 times

[Tubetec]

Ok ,
I figured something out ,
when testing with REW I obviously wasnt applying phantom power into the headphone amp on my interface , Id left the battery out and the power switch to off , that bypasses a set of caps and two resistors in the signal path , doh !
Explains a lot ! ,

All the test results were flawed because of opperational errors on my part ,
I'll need to re run the curves again ,

I also made up a simple load adjustment , it consists of a male and female xlr socket, a strapped 25kpot and a 1k resistor in series across the load ,
Im able to vary input impedence from around 1k to 12.5k on line input ,giving very usable shift in frequency response from the basic 4 positions ,
I did a sound test with the rode NT1a , results were good , you can counteract proximity effect precisely at any distance .

[mediatechnology]

What is the value of C3 and C4?

[Tubetec]

47Uf-50v Electrolytic ,
10 ohm resistor in series ,

Two graphs ,
first one with 1.2k load , second 12.5k load

[mediatechnology]

The largest measurable difference with the (47 uF X2) in circuit will be at Cut 1 which will raise Fc by about 20%.

[AnalogJoe]

I am definitely not a fan of electrolytics for filters, they have very high tolerances and will most likely unbalance the cut-off frequencies and phase vs frequency between the Hot and Cold wires. In some cases this could impair CMRR.

[mediatechnology]

I am definitely not a fan of electrolytics for filters, they have very high tolerances and will most likely unbalance the cut-off frequencies and phase vs frequency between the Hot and Cold wires. In some cases this could impair CMRR.

I did quite a bit of tests on electrolytics and found that caps from the same lot usually ratio match within 1-2%. This post is just a few of the various lots I tested: viewtopic.php?t=1234

Did more tests for THAT but haven't published them.

Also checked some NOS 100 uF bipolar (MCI stock) from the 80s and they closely matched too and were surprisingly close to original value.

But for HP filtering they are not the best.

[AnalogJoe]

I am definitely not a fan of electrolytics for filters, they have very high tolerances and will most likely unbalance the cut-off frequencies and phase vs frequency between the Hot and Cold wires. In some cases this could impair CMRR.

I did quite a bit of tests on electrolytics and found that caps from the same lot usually match within 1-2%. I have a post on that somewhere I'll link to.

But for HP filtering they are not the best.

I have no problems with electrolytics as decoupling caps, provided they are large enough, but there are several issues why I do not like electrolytics in filters. Tolerance is one of them. Low self-resonance frequency is another, which is related to what you mentioned about HP filters. Matching is also another issue when trying to implement filters of 2nd or higher order, which can severely alter the filter response, especially if you are using high-Q filters. Dielectric absorption might also become an issue.

Electrolytics distort even when a small signal voltage drop is impressed across them, which makes them perfectly ok as decoupling caps if you place the cut-off frequency so low that the voltage drop across the cap is insignificant at audible frequencies, but if you use it as a filter with a higher cut-off frequency, they will have a significant voltage drop across them at audible frequencies.

I took a look at the original cap values TubeTech posted. I believe you can find all of them in a poly-type cap. The 12 uF cap is probably very bulky (and not cheap) https://www.mouser.com/ProductDetail/Pa ... ZIfQ%3D%3D, but still manageable, or several small caps could be placed in parallel.

[mediatechnology]

Electrolytics distort even when a small signal voltage drop is impressed across them, which makes them perfectly ok as decoupling caps if you place the cut-off frequency so low that the voltage drop across the cap is insignificant at audible frequencies, but if you use it as a filter with a higher cut-off frequency, they will have a significant voltage drop across them at audible frequencies.

I agree but don't you mean to say "coupling" (DC-blocking) rather than "decoupling?" (supply bypass)



I like the first bootstrapped FET input circuit better as it uses small value films.

[Tubetec]

Wima actually do a 47uf 50v MKS foil ,
So its definately possible to do , but would cost around 50 euros in capacitors .

The SSL2 has balanced hi-z fet input , so I could try making a version with smaller foil caps ,
Might be useful to try a version dedicated to line input(10-20kohms) also.