KM84 circuit experimenting in progress ...

[HellfireStudios]

The break will do you you some good. Sometimes I wait for up to a week before I retackle a project. The situation is similar to the ear fatigue you can experience when mixing for too long...

-James-

 

[MagnetoSound]

I'm pretty sure you have the drain and source reversed.

There are a few conflicting pinouts on the web, but according to Fairchild if the flat is facing upwards and pins are towards you, the left-hand pin (pin 1) is the drain.

 

[gary o]

Thanks Dan I came across that data sheet & as you say others and as you D & S other way round ......I have had my FET wired opposite to this fairchild data sheet but did reverse it in the circuit I posted on here that it seemed to make no difference ? even the voltages were still right ?

Makes me want to give it another go with another FET..... I put it aside but might warm up the iron .....

cheers

 

[Rossi]

In my experience, reversing D and S makes almost no difference. I do have a look at the datasheet every time I use a new FET type. Also, there some FETs (e.g. BF types) that do not have the gate in center position.

 

[volker]

In my experience, reversing D and S makes almost no difference. I do have a look at the datasheet every time I use a new FET type. Also, there some FETs (e.g. BF types) that do not have the gate in center position.

There are many FETs that are quite symmetrical, so D and S are interchangeable. That is often stated in the datasheet, too. But for some it doesn't work all that well.

 

[gary o]

I just did this again.....

Put your DMM in current measuring mode.  Take a Jfet, short gate and source together, and connect to negative terminal of a 9V battery.  Connect one DMM probe to the drain, the other to the positive 9V battery terminal.  You should be able to read the IDSS on the meter.

If your meter can't measure current, then put a 10 ohm resistor (measure it first!) between the drain and the positive battery terminal, then read the voltage drop across the resistor and divide the voltage by the resistance you measured.
[/quote


but put source & gate round the other way ...... got same measurements kinda supprting that it may not matter which way round gate & source are.... but can that really be true....why have a gate & a source if so ..... but yeah in my circuit it didnt seem to make a difference ......

And now the gate might not be middle leg nooooo .....

I had two packeds of FETS that I have lost track of where they came from ....at least one cmae from Farnell recently ...... other prob Farnell too but earlier date ..... I have two FETs I found in my junk box unused one looks much older was saving it in case it had special vintage pixy dust on it ..... but I think its just moldy from being in the shed for 20 years...

Got fed up with the KM last night & had play with Oktava 012 circuit sounds great was tweaking bias trimmer when little bit of back ground noise dissappeared completly its totally silent , I love that circuit ....

cheers

 

[HellfireStudios]

If changing the FET has no affect on the problem, it may not be the problem.

-James-

 

[MagnetoSound]

... but put source & gate round the other way ...... got same measurements kinda supprting that it may not matter which way round gate & source are.... but can that really be true....why have a gate & a source if so ..... but yeah in my circuit it didnt seem to make a difference ......

And now the gate might not be middle leg nooooo .....

I had two packeds of FETS that I have lost track of where they came from ....

Source and drain ... !

I have been ASSuming all along that you have been using a 2N3819, as in the original circuit - in which case, the gate is definitely in the middle.

Even so, there is a caveat, as not all manufacturers follow the same convention with the pinout. It is worth checking the data sheet from the actual manufacturer of the specific 3819 (or other FET) that you are using that the pinout is accurate. I have seen one 3819 data sheet where the source is on pin 1.

As Volker says, some FETs are more symmetrical than others - but also, small signal circuits like this where bias setting is critical are much more sensitive to it, and noise and/or loss of gain can occur if the bias is inverted, IME.

Of course, if you happen to have picked a FET where the gate is to one side, obviously all bets are off! ;D

 

[gary o]

Its been a while but I had another look at theis experiment .... At first as before it didnt work very well very low signal my voltages seemed ok but couldnt see 46V on R1, it did read 41mv tho .....I thought at my previous attempt I had 46V at that node ..... or maybe I had 46mv .... I should be able to measure with a nomal digital meter 46v at that node right ?

Thing is I removed C6 & now I have a very decent sound level ..... I added a home made 4pf C2 feedback cap, level went down a little, I made C3 22uf for now for more bass & that worked now I have very nice sounding mic.... Im using a T14/1 transformer 11:1 ratio (I think) using Violet Vin67 capsule.

I have the 10k trimmer + 3K res to ground on the source I can trim to get 10V on the drain, the voltage on the source is about 2.2V.

As it it is it sounds pretty bloody good .... I dont understand whats going on with the R1 R6 node as I dont have 46V it measures about 8V ? & as I say I have to disconnect C6 ? & still not sure the source voltage is ok ....

Im going to tidy up the circuit & maybe tweak a little ...

 

[MatthisD]

Are you sure C6 is causing a problem? without C6, R8 and R5 is a 20M resistor. If you have the right values of 10M/100nF then I can't see how it would drop the gain. You can't properly read the capsule voltage there with a meter because of the volts dropped across 20Meg of resistance once the meter probe makes contact. Tip your probe there for a split second and see what your meter reads.

 

[gary o]

Are you sure C6 is causing a problem? without C6, R8 and R5 is a 20M resistor. If you have the right values of 10M/100nF then I can't see how it would drop the gain. You can't properly read the capsule voltage there with a meter because of the volts dropped across 20Meg of resistance once the meter probe makes contact. Tip your probe there for a split second and see what your meter reads.

No not sure thats why I ask  ..... does removing C6 0.1u cap upset the voltage at R1 R8 node....... So a cheap digital meter would NOT show 46V at R1 R8 not anyway.... it would upset the circuit..... with no C6 my meter reads 8V, not saying thats right...with the dodgy C6 ( explain in mo) the R1 R6 node read 46mv by chance...... back further in this thread I reported that it read 46V I wonder in my haste if it back then read 46mv.....

The dodgy C6 it doesnt short but when I connect my new capacitance meter (only cheap but fun) the display flashes & wont display a reading......I tested C10 with meter it read 0.1 uf right away....

Something I didnt do was match R9 R10 1%.... I will do so....

For the moment I have housed the electronics minus the C6 in a round tin & have lollipop head on top even as it is it is more than useable slightly less level than a Oktava MK319 sound is kinda soft but not bad at all .....I still have wrong ratio transformer in there too .....

 

[HellfireStudios]

Did you bias the FET? The bias, much like the U-87, plays a major role in the amount of unclipped signal that will pass through it (this translates to higher gain and lower S/N ratio). I perfboarded a pair KM-84 circuits into some MXL 604's. Mine have been working quite well so far. The transformer I found was a Cinemag CM-2510 (8:1). I hooe you get it working properly soon. It is a great circuit.

-James-

 

[gary o]

Thankyou James  I think Im nearly there, I tested my FETs with multimeter .7ma was my best I have a 3K res from ground & 10K trimmer trmmed to 10V on the drain I got 2.2V on the source now would like to get lower if possible but it does sound pretty good and noise is pretty good too....

Update from this morning I replaced my C6 now & mic still sound good so I assume my meter gives false reading of 8 volts at R1 R8 node...

Im going to try a OEP transformer on output.

Im also thinking about sendng polorisation of 60 Volts from another power supply into the mic.... hoping not to harm capsule or FET..... Im not sure but I think if I disconnect  R1 from R8 & inject the 60 VDC R8 side of R1 it should work ?

 

[micaddict]

KM74?

 

[gary o]

KM74?

??

 

[micaddict]

Neumann KM74 was the precursor of the K84.
Unlike the KM54 and KM64, it did not have a tube, but it still did have a PSU.

 

[gyraf]

KM74 did not really depend on an external PSU - there are no separate power input pins on the connector - it was made for 12V "TonaderSpeisung", a kind of 12V-phantom from the time before P48 became a standard..

Jakob E.

 

[micaddict]

I stand corrected then. Thanks, Jakob. 

BTW, ever heard of this guy?
http://www.rensheijnis.com/microphones.htm

Scroll to "Separated main power supply for microphones".

 

[Steph.B]

Neumann KM74 was the precursor of the K84.
Unlike the KM54 and KM64, it did not have a tube, but it still did have a PSU.

No, not really.

The Neumann KM74 and KM84 were out at the same time, in 1966 precisely.

The KM74 was part of the fet70 serie, comprising the KM73-74-75-76 and the U77.
They didn't need an external supply but worked with 12V T-power, you probably mistake the fact that most of these microphones now shows with a box converting 48V phantom power to T12.

The KM84 was part of the fet 80 serie, which included the KM83-84-85-86 and the U87.
They of course work with 48V phantom power.

One interesting fact is that the fet70 serie microphones are all transformerless. I think they are Neumann's firsts (with their ancestor/prototype the KTM) transformerless microphones before the TLM serie.

 

[micaddict]

Yes, point taken. Apparently it was too long ago for my dusty brain. I should have checked.

Nice discussion, though.