Jensen Cascode Mic Pre

[raysolinski]

Anybody here build this pre? It is on their site under schematics. I just noticed that I have all the parts for this guy layin' around on the bench. Well, except the input transformer but i think I will order a Jensen to keep it in line with their design..plus i don't have a pre with any Jensen's, just Cinemags and Lundahls.

Cheers,
Ray

 

[skipwave]

I'm intrigued by cascode topology. It's like an exotic food I want to sample, but can't quite justify a whole outing for.

This one (as054) has an interesting note about direct coupling to a 12BH7 output driver stage, shown in as021.

I just pasted the two together, and it looks like a drunken "Ring of Three" seeing all those tube symbols in double.

 

[cory ponzo]

Haha years later, I built this preamp in 2024. Parts from a Akai reel to reel. Edcor Transformers. And Condor 48 volt linear supply for phantom. Has ez81 rectifier dc filaments. And a power supply from a 5f1 Champ.

Sounds amazing. I'm going to build more. My first one isn't the most attractive. The bass DI is Soooo good. I used 1:6 input transformer. 1:4 output. So I used two 12au7 as it had to much gain With v1 12ax7 and was motorboating.

I learned so much. To make it as quiet as possible I tied all grounds to the power transformer base, all except the tube power supply ground and the in/out shield. Also note the orientation of the in and out transformer. It is the lowest noise tube pre I have used.

Definitely build one!

 

[rock soderstrom]

Haha years later, I built this preamp in 2024. Parts from a Akai reel to reel. Edcor Transformers. And Condor 48 volt linear supply for phantom. Has ez81 rectifier dc filaments. And a power supply from a 5f1 Champ.

Sounds amazing. I'm going to build more. My first one isn't the most attractive. The bass DI is Soooo good. I used 1:6 input transformer. 1:4 output. So I used two 12au7 as it had to much gain With v1 12ax7 and was motorboating.

I learned so much. To make it as quiet as possible I tied all grounds to the power transformer base, all except the tube power supply ground and the in/out shield. Also note the orientation of the in and out transformer. It is the lowest noise tube pre I have used.

Definitely build one!

Thanks for info, did you change anything for the 12AU7? Here is the schematic:


PS: the second gain stage is not a "cathode follower" as noted in the screenshot above, it is a SRPP...

Edit: here the original schematic from the Jensen website

 

[cory ponzo]

Haha years later, I built this preamp in 2024. Parts from a Akai reel to reel. Edcor Transformers. And Condor 48 volt linear supply for phantom. Has ez81 rectifier dc filaments. And a power supply from a 5f1 Champ.

Sounds amazing. I'm going to build more. My first one isn't the most attractive. The bass DI is Soooo good. I used 1:6 input transformer. 1:4 output. So I used two 12au7 as it had to much gain With v1 12ax7 and was motorboating.

I learned so much. To make it as quiet as possible I tied all grounds to the power transformer base, all except the tube power supply ground and the in/out shield. Also note the orientation of the in and out transformer. It is the lowest noise tube pre I have used.

Definitely build one!

Oh some pics... lol not the most elegant. I wasn't sure about layout and how to run wires for low noise.

 

[cory ponzo]

Thanks for info, did you change anything for the 12AU7? Here is the schematic:
View attachment 139199

PS: the second gain stage is not a "cathode follower" as noted in the screenshot above, it is a SRPP...

No changes it was motorboating around 12 o'clock. I read in another forum motorboating could be from too much gain. I tried a 12au7 and all was good. So I left it.

 

[rock soderstrom]

Oh some pics... lol not the most elegant. I wasn't sure about layout and how to run wires for low noise.

I don't see a PE pole, is your box connected to the protective ground?

It was certainly not easy to get the amp quiet with the unshielded Edcor transformers, especially the input TX is a challenge with all the gain afterwards!

 

[rock soderstrom]

So I used two 12au7 as it had to much gain With v1 12ax7 and was motorboating.

Motorboating can be caused by the wiring; you may be able to avoid this in your next amp by optimizing it. Keep the inputs of the gain stages away from the outputs. A Mumetal shielded input transformer can also help and would simplify the setup considerably.

 

[Alva Goldbook]

No changes it was motorboating around 12 o'clock. I read in another forum motorboating could be from too much gain. I tried a 12au7 and all was good. So I left it.

motorboating is typically caused by insufficient B+ filtering, or it can be due to low frequency oscillation. All wires emit magnetic fields, so keeping them short and tidy is always a good idea. High current wires like heater supplies should be kept away from sensitive grid wires. In this kind of a circuit, heater elevation would probably be a good idea. This means referencing the heaters at some arbitrary higher DC voltage like 30VDC instead of referencing them to ground (0 volts)

 

[pucho812]

Motorboating could be a polorized cap in backwards that is rated higher voltage then what is passing through it.

 

[MisterCMRR]

Motorboating could be a polorized cap in backwards that is rated higher voltage then what is passing through it.

Motorboating is nearly always due to insufficient decoupling between the B+ for output tube and input stage. One solution is a voltage regulator, dropping B+ by 10 or 20 V. Alternatively, substantially increasing capacitance on the rail will "brute force" it to be stiffer. Magnetic coupling between wires running parallel is always worse at high frequency. If filaments operate on DC, magnetic coupling is virtually zero (only changing magnetic fields couple to induce voltage into nearby wires), but twisting of filament pairs is always a good practice.
As it's designer, I'm pleased to hear praise about the circuit's sound. I'm very fond of cascade configurations, whether solid-state or vacuum-state!

 

[MisterCMRR]

Make that cascode circuits (spell checker not good in technical contexts!!)

 

[Alva Goldbook]

Make that cascode circuits (spell checker not good in technical contexts!!)

I've been fascinated about cascode circuits for a while. Never gotten the chance to experiment with one though. From what I've read the 12au7 is particularly well suited for such circuits.

 

[ruffrecords]

The Jensen design has no separate decoupling of the HT supply to the first stage. This means that a poor layout could easily lead to current changes in the output stage finding their way back into the input stage. I would suggest you try a completely separate HT and 0V connection from the first stage right back to the PSU, do not go via any part of the second stage and do not collect $200.

If this does not work, insert a 10K resistor in the HT feed to the first stage and fit a 22uF 300V electrolytic from the junction of R56/R8 to 0V.

Cheers

Ian

 

[Alva Goldbook]

The Jensen design has no separate decoupling of the HT supply to the first stage. This means that a poor layout could easily lead to current changes in the output stage finding their way back into the input stage. I would suggest you try a completely separate HT and 0V connection from the first stage right back to the PSU, do not go via any part of the second stage and do not collect $200.

If this does not work, insert a 10K resistor in the HT feed to the first stage and fit a 22uF 300V electrolytic from the junction of R56/R8 to 0V.

Cheers

Ian

Good catch. They should be decoupled from one another. Reservoir cap should be around 100uF from my calculations.

 

[ruffrecords]

Good catch. They should be decoupled from one another. Reservoir cap should be around 100uF from my calculations.

Care to share?

Cheers

Ian

 

[MisterCMRR]

Since these were two separate application schematics, details of combining them are left to the reader as an exercise. Had it been a schematic of a complete preamp, I would have provided at least one decoupling method - my first choice being a voltage regulator like an LM317. Yes, they can be used to regulate high voltages since the regulator only "sees" the differential voltage across its IN to OUT terminals and its ADJ terminal sits 1.25 V below the OUT terminal.

 

[Alva Goldbook]

Care to share?

Cheers

Ian

Sure, I don't mind sharing the math.

So for DC ripple here's the formula, with many thanks to Merlin.

C = I / (2 f Vripple)

We got a 12ax7 and a 12au7. Typical current draw would be 10mA per 12au7 triode and 1mA per 12ax7 triode. These are ball park figures, but should be good enough for the calculation. Which gives us a total of 22mA. So I = 0.022A

Assuming we're dealing with 50Hz mains frequency, and a B+ of 250v, and we want to reduce ripple down to 1%, we get, 2.5v x (50x2), which neatly gives us 250v. So if we want to shoot for 1% and we're dealing with 50Hz mains frequency, we can simplify this to

(2f Vripple) = B+ voltage.

0.022/250=0.000088 Farads. Now that's too many zeros not to make my dyslexia go bananas, so I just do the farad to uF conversion with a online calculator. But it ends up being 88uF. Which isn't a standard value, and a little extra doesn't hurt. So 100uF.

 

[MisterCMRR]

Note that I put DC operating voltages at key points in the schematics. Using these and Ohm's law, you can see that the plate current in V1 is 1 mA, in the voltage divider is 0.36 mA, and in V2 is 2.5 mA. So, if we insert a 10 kΩ series resistor in the B+ line to the V1 stage (1.35 mA total), it will reduce B+ to that stage by about 13.5 V, which is inconsequential. Adding a bypass capacitor at the V1 side will reduce existing ripple by approximately the ratio of 10 kΩ to the capacitive reactance of the C. At 120 Hz (full-wave ripple of 60 Hz supply), the reactance of 20 µF is 67 Ω. This will give 43 dB ripple reduction (to about 1% of its original value). For 50 Hz supply, it will be about 1.5 dB less. This filter should also have enough attenuation at sub-sonic frequencies to stop any "motor-boat" oscillation