Jfet pre

[thor.zmt]

Sorry @thor.zmt bit is this schematic for your tests or do I install it as it is ? [10Ω > 1mH] + [10mF + 10MΩ to ground] ?

Every electronic component has unwanted behaviours, commonly we characterise and model them as "lumped" model, meaning a capacitor for example looks like this as a real part and what we can use for a first order approximation:



Most simulators come with "Ideal" passive parts, to get simulations that make sense the operator must either add the extra parts to make the simulation workable or create complex models.

For Audio a first order approximation usually is fine, so I add the resistance of the choke as discrete resistor and the ESR of the Capacitor(s) to get a realistic result.

And I will change my 3,9KΩ by 3,3KΩ (there is 100Ω is before every switch - the actual configuration is : smps 48v > 100Ω > switch + 100uF • 63v > 3,9KΩ > Tr center tap). The other solution is to put 6,8KΩ on Pin2 & 3 and forget the center tap (?)

Up to you. You need to evaluate what will crate the worst situation for net DC current differences.

Considering what I just learned about input transformers, may I ask you one more advice please ? in my quad pre's I use CAPI CA2622 with (again !) 220KΩ secondary charge...
- does it cause any problem ?

It has the potential to cause problems. It is impossible to know without test gear and performing tests.



(this omits the resistance's, imagine a resistor in series with each capacitor and inductor)

- what value should I install ?

Usually transformers are designed to give a more or less flat response for a specific set of loading and source impedance's. If you want the specified result, use the manufacturers specified circuit.

- what would be the benefit ?

Depends what you want. If you want a flat frequency response without peaks and / or excessive roll-off, stick what these Transformers were designed to use.

In my tube circuits (12AY7 + 12AU7 AF) I didn't install any gain level (I think I'm gonna change this) but a 100KΩ pot. Log (with 150pF from in to out) after the 12AU7 and then a cascade jfets Z adaptation. It sounds very well.

You did not measure frequency response or distortion, did you? Given how inexpensive high quality and flexible test gear is these, this is easily changed.

Thor

PS, that "spec sheet" you posted lacks any information that would allow the customer to determine how to actually apply this transformer, unless the customer does the whole work of characterising the transformer in detail.

Please see this for comparison:

https://www.jensen-transformers.com/wp-content/uploads/2014/08/jt-110k-hpc.pdf

That is a "spec sheet".

PPS, while "sounds good" is the goal, one must be careful to not be deceived... See:

https://tu-dresden.de/mn/psychologie/ifap/kknw/die-professur/news/we-hear-what-we-expect-to-hear

 

[Emmathom]

I will ask CAPI for further informations about their transformers.
I understand now that the 10Ω & 10MΩ (at PSU) were put to simulate real situation.
Forgive me, but as you probably noticed, I'm far from experimented : I began to seriously to get involved in electronics just 2 years ago.

So I'm gonna order the components I miss and will get back to you @thor.zmt when I'll be ready.
Thanks +++ for your help & have a good evening

 

[Wavelength]

Why no cascodes or good sounding JFETS like the LSK170 or the Toshiba version 2SK170. If you want more input range use the LSK189 or LSK489 dual. Going to be a lot of input capacitance with =out doing a cascode. I use a folded cascode a lot in pedals and some preamps I have made. double the swing voltage than the supply voltage. The J113 is ok, not great sounding I use it in power supplies some times for shunt designs. https://www.linearsystems.com/

 

[Emmathom]

Hi @Wavelength
Thank you for your suggestion.
It is very pleasant to see that you are offered help on this forum when you request it, and several members have commented on this thread with a positive spirit... Nice ! but the problem, when you receive several proposals, is about making a choice without hurting anyone. I am thinking in particular of @thor.zmt who invested himself in this project with seriousness and concern to meet my expectations. I thank him once again.

But in the end, if I have several choices offered to me, I must decide according to what suits me best.

I don't know if you have read this entire topic and thus understood what my objective is but I will make a recall here. Some asked me why I didn't build a tube circuit: I already have 6 tube channels (listed in 2 devices) which give me complete satisfaction and for the preamp we're talking about, a 4 channels "mixed" tubes + transistors, 1) I cannot add 2 additional tube channels without having to make major modifications. Additionally, my case is relatively small... 2) I like the idea of having 2 different sounds (or technology) in the same gear (can be appreciable in some recording situations).

So, I want to swap the 2 BJT channels currently installed by jfets (for reasons argued earlier). If I wanted just a "clean" sound without "personnality", I would have opted for AOP. I know how to deal with without asking for help since I built a quad AOP 2 years ago / my first preamp project and it works... fine. Not much more to say.
So if no tubes, no BJT, no AOP, my choice naturally falls on jfets which, in addition, seem to match best with the 2 tube channels already installed.

I have not yet ordered the missing components allowing me to assemble the Thor diagram because I don't have to rush (I have no emergency at the moment since I'm working on diy microphones).

I am therefore open to any proposal, remembering that I like simple and effective things. I have no need for a phase inverter, an impedance selector, etc : I want a preamp suitable for condenser microphones (rather very sensitive and generally 200Ω output Z) with, if possible, a single gain potentiometer and that's all. I don't intend to buy extra output transformers and don't need a symetrical output. In my pre's I use a simple jfet cascade referred to N. Pass and it does the job great.
I'm a fan of this quote: the best part in an audio chain is... the one you can do without!

If you have a diagram to submit to me, I am of course interested and I thank you in advance.
PS : I have in stock 4 (only 4) true 2SK170GRs with very close IDSS and VgsOff just waiting to work (but today I would buy the smd version ! > 0,45€ each and more technical advantages)
Best Regards

 

[thor.zmt]

Why no cascodes

Simplicity.

or good sounding JFETS like the LSK170 or the Toshiba version 2SK170.

The J113 sound very good, I used the SMT version in many products. Just because it is an American JFET doesn't make it bad. It is quite comparable to 2SK170 except noise is around 6-10dB higher which in many cases is of little consequence.

In our case, 2SK170 et al. have too little bias.

LSK498 etc... Maybe. But the goal was cheap and simple.

But then, why not J111 and redesign?

Going to be a lot of input capacitance with =out doing a cascode.

Really?

The ~12pF Ciss at the operating conditions is basically bootstrapped by the un-bypassed source resistor and thus reduced effctively to less than 5pF.

Crss is ~5pF. The Gain of the first stage is 12dB to guarantee overload margins.

So effective Cin for the first stage is ~25pF. I think this compares well to many Tube circuits and will be fine.

The second stage supplies the bulk of the gain, but it's still only ~30dB. So we get around 160pF effective input capacitance for a simple common source amplifier. Again, not terrible next to (say) ECC83.

For a number of reasons (the Mu-Follower among others) in reality we get less effective Cin in the real design, around 100pF, instead of 160pF.

In combination with the 50k Gain control, in the worst case for this circuit, we get a 32kHz -3dB Point, at lower gain settings things rapidly, for -6dB gain on the dial we get 63kHz -3dB.

Given the bandwidth constraints of the transformers used I am not sure I am unduly worried over the bandwidth constraints of the deliberately simple active circuit.

I use a folded cascode a lot

So do I. And other options

But the requirements were:

very simple circuit
"Tube style" characteristics, specifically, this:

https://www.tangible-technology.com/tubes/1566/1566.html

To talk about the "sound" of a part, even a transformer, without circuit context is misleading.

A Y5U ceramic capacitor is terrible sounding, right?

In my BoomBoxx (aka Beach Blaster) build I deliberately used a pair of Y5U DC blockers (blocking ~12V DC at the input of the Class D Amplifier) and I deliberately undersized them with a ~70Hz cut-off.

And guess what, it sounds MUCH better with Y5U compared to film. Subjective Bass response is significantly improved, especially when playing loud, like on the beach (for quiet there is a Bass Boost).

The BoomBoxx has a pair of cheap 6.5" 2-Way car kit speakers from the ex. OEM of Focal/JML. Vented box, 42Hz EBS tuning. 92dB/1m @ 2.83V per driver, so in stererererereo 98dB/1m @2.83V * 2 and with +15dBW output per channel we end up with 113dB peak / 1m.

Something with this level of output and man (or tiny woman) portable will never be a bass monster no matter what. By using a capacitor to deliberately create a lot of H2 at low frequencies, on a commercial Class D Amplifier board, subjective bass response is much improved (this is similar to "MaxxBass" and is based on the "missing fundamental" effect).

So it's always a question of context. My experience is that the J113 and MMBFJ113 On Semi only are often an excellent replacement for a 2SK170, if small circuit adjustments are made. Only where very low noise is a concern is a 2SK170 better, or a LSK170.

Thor

 

[Emmathom]

Hi @thor.zmt
If we take a great advantage at buying better jfet ref. (in Process 55), do not hesitate to tell me.

About purschasing components : always a pain in the ass to get all you need at the same retailer... (shipping costs)
I found the 1mH Abracon self and the capacitors at Mouser's (I got all resistors > Vishay Dale 1/4w for R11 / R14 / R2 / R4 / R27 - others are generic metal 1% 1/4w except for the R6 2x2,2KΩ in // 1w metal 1%)
BR

 

[thor.zmt]

If we take a great advantage at buying better jfet ref. (in Process 55), do not hesitate to tell me.

I do not like process 55 parts much for audio, process 51 is better.

About purschasing components : always a pain in the ass to get all you need at the same retailer... (shipping costs)

If I order Nichicon Muse Cap's at Mouser I commonly get free shipping to Thailand....

I found the 1mH Abracon self and the capacitors at Mouser's (I got all resistors > Vishay Dale 1/4w for R11 / R14 / R2 / R4 / R27 - others are generic metal 1% 1/4w except for the R6 2x2,2KΩ in // 1w metal 1%)
BR

Yes Vishay Dale "Brown Devils" are a good choice.

Thor

 

[Emmathom]

Answer from CAPI about the CA2622 transformer as I asked for further informations...

Hello

In a 312 card there was no load R. In most other preamps and in the console a 150k load R was used. We use a 10k -> 220pF as a zobel.

Best regards, Jeff Steiger
Classic Audio Products, Inc.
http://www.capi-gear.com
https://www.instagram.com/capigear/
https://facebook.com/jeffsteiger.capi

 

[Emmathom]

PPS, while "sounds good" is the goal, one must be careful to not be deceived... See:

https://tu-dresden.de/mn/psychologie/ifap/kknw/die-professur/news/we-hear-what-we-expect-to-hear

Very intersting indeed... how our brain can be (is) manipulated by so much external datas !

 

[thor.zmt]

In a 312 card there was no load R. In most other preamps and in the console a 150k load R was used. We use a 10k -> 220pF as a zobel.

So the optimum load for damping HF response peaks is 10kOhm. Using a Zobel reduces the load, but you need sufficient test gear, or a reliable recommendation from the Manufacturer for the RC snubber/Zobel.

API 312 actually shows 220pF/5.1kOhm (R4, C5)!



Thor

 

[Emmathom]

The datasheet of CAPI CA2622 is false : the output is wired with to windings... and this is how mine are implemented.
Now the modification is to "charge" the output, not with 220KΩ but with 50KΩ.
According to what is said (?) Primary in // = 150Ω and secondary in serie = 10KΩ so √(10000/150) = 1:8,17