Lorenz Mic-Pre

[DaveP]

I came across this in my files last week and thought it needed an airing.

It's the East German version of the V76 mic-pre and it has some interesting features.

I've cleaned up the schematic and translated it but I've left the stuff in the drawing box as is.  The original is attached below.

What I found interesting was that there was anode to anode feedback.  This increases the anode current and makes the ratio between the anode current and the g2 current, much greater than it would normally have been which is a clever way to reduce partition noise.

The other clever idea is the feedback of a small portion of the heater voltage for hum cancellation.

The dotted lines around the transformers and part of the circuit indicate shielding encasement.

It's dated the 18th Sept 1957, so it's contemporary with the V series mic pres made in west Germany.

I have no specs on the amp but I'd like to hear what others think of it

Best
DaveP

 

[abbey road d enfer]

What I found interesting was that there was anode to anode feedback.

This is really anode to grid NFB, both of the 2nd tube. The fact that the grid receives signal from the anode of the 1st tube does not make it part of the NFB, although its output impedance is an important parameter.

This increases the anode current and makes the ratio between the anode current and the g2 current, much greater than it would normally have been which is a clever way to reduce partition noise.

Are you sure? Anode current of the 1st tube is constituted of 0.46mA from the anode res and 0.16mA from the FB res. I would say the governing factor in the difference between anode and g2 current is predominantly the high value of the g2 res (1.6Meg). With about 0.11mA g2 current, the ratio between anode and g2 current is not too dissimilar to the more common applications. Anyway, the EF86 operates pretty far from the recommended values; running it at about 0.5mA makes it about 4dB noisier than in a REDD47, where it runs at about 1.5 mA. In addition, the 16k cathode resistor adds non-negligible noise. However the 1:18 input xfmr tends to compensate, with nearly 6dB "free-gain" than most comparable designs.

The other clever idea is the feedback of a small portion of the heater voltage for hum cancellation.

That is a nice touch. It is a good palliative to the use of DC heaters.

It would be interesting to do measurements in order to see if any spectacular performance aspect justifies the quirkiness of the design.

 

[ruffrecords]

This is an interesting design. Elements of the V76 and the REDD47. It does not fully implement the V76 technique of varying open loop gain in step with closed loop gain to ensure stability over a wide gain range and it does not quite balance the operating points of the two tubes to help cancel distortion as in the REDD47 design With some simple changes it could make a really good preamp.

Cheers

Ian

 

[DaveP]

The 2M grid resistor on V1 adds much more noise than the cathode resistor me thinks.

 

[abbey road d enfer]

OOps! Sorry, Dave, I hit the wrong button.  I erased most of your post.
However, here's my answer about the 2 Meg resistor.
It's the equivalent resistance that counts. The actual Johnson noise source is the reflected resistance of the source + DCR, which account for about 70kohms. The 2Meg is negligible.

 

[rock soderstrom]

Interesting topic, technically and historically.

"in Verwaltung" is a nice expression for forced nationalization (or socialization?? )

 

[DaveP]

That's interesting, I never bothered translating that part.

I like translating old German technical stuff, I find the meanings have changed a little over the last 60 years though

DaveP

 

[rock soderstrom]

Here is a link to an articel of a german journal called "Studio Magazin". The have released a review of some microphones. In this review you will find a (german) description and some measurements of the Lorenz amp. (page 7/8, quite blurry...)

http://echoschall.de/downloads/Studio_Magazin-Mikrofontest-19.pdf

 

[alexc]

Most interesting ...  I thought it was anode-to-anode nfb  as well as anode-to-grid nfb on the second stage.

Kind of a 'schade' style nfb over both stages, coupled to an anode follower output stage.

Just working thru the detailed explainations   Still don't understand thecathoder circuit gain switching of the first stage.

Very interesting and another classic circuit to study some more  ....

Thanx

 

[abbey road d enfer]

I thought it was anode-to-anode nfb  as well as anode-to-grid nfb on the second stage.

NFB is by definition something that happens between an output and an input. The fact that this input is connected to something else doesn't change this predicament. The properties of the signal source matter, though, particularly in this case of voltage-to-current NFB, where NFB actually decreases the input impedance of the stage it is applied to.

  Still don't understand thecathoder circuit gain switching of the first stage.

Maybe it's because of the way Dave redrew the schemo; look at the original one. Switching is not very difficult to grab.

 

[ruffrecords]

Most interesting ...  I thought it was anode-to-anode nfb  as well as anode-to-grid nfb on the second stage.

Kind of a 'schade' style nfb over both stages, coupled to an anode follower output stage.

It is what is commonly called Schade NFB but why anyone needs to give it a special name I do not know. Think of the second triode as an op amp. The anode is the output and the grid is the -ve input. So there is NFB from op to -ve input so we have a virtual earth at the grid (not a very goo one because the triode gain is small). The gain of the triode stage is defined by the NFB arm divided by the the input arm but it appears we do not have an input arm because it is connected to the pentode plate. So the input arm is actually the output impedance of the pentode section. This is the pentode's ra in parallel with its plate resistor and since the pentode's ra is very high we can approximate it as the plate resistor so the triode stage gain is about 1Meg divided by 330K

Just working thru the detailed explainations   Still don't understand the cathoder circuit gain switching of the first stage.

Very interesting and another classic circuit to study some more  ....

Thanx

This is a variation on the V76. The pentode stage gain is approximately its plate resistor divided by is cathode resistance. If you keep the open loop gain constant, and reduce overall gain by increasing the amount of feedback, you get instability so the V76 trick is to simultaneously reduce open loop gain  as the amount of feedback increases by increasing the pentode cathode resistance so as to reduce its gain and maintain stability. In this design this is achieved only at the two lowest gain settings where additional resistance is added in series with the pentode cathode.

Cheers

Ian

 

[alexc]

Thanks Ian, that clears some things up for me   Some good detail there in your description, for sure.

The 'schade' thing comes mostly from the 'other' tube forum, I think

http://www.diyaudio.com/forums/tubes-valves

They reference that classic paper (on nfb in tube stages) quite a bit with respect to various topologies,  so the terminology is thrown about readily by some, including myself

 

[abbey road d enfer]

Kind of a 'schade' style nfb over both stages, coupled to an anode follower output stage.

First time I hear about Schade NFB and about the O H Schade character. A matter of cultural differences...
Here its local voltage-to-current NFB.

 

[ruffrecords]

First time I hear about Schade NFB and about the O H Schade character. A matter of cultural differences...
Here its local voltage-to-current NFB.

There are too many ways of describing negative feedback, many of which are inaccurate and some of which are downright misleading (like Schade or anode to anode). My personal preference is to state how the feedback is derived and applied - there is no need to refer to voltage or current. In this case, the NFB is shunt derived and shunt applied. I prefer this description because it immediately tells you two important effects of the NFB. Any feedback that is shunt derived lowers the output impedance, Any NFB that is shunt applied lowers input impedance. The only other way to derive and apply NFB is in series and series derived NFB increases output impedance and derived applied NFB increases input impedance. So,

shunt derived NFB reduces output impedance
Series derived NFB increases output
Shunt applied NFB reduces input impedance
Series applied NFB increases input impedance.

Cheers

Ian

 

[DaveP]

Returning to the circuit,  What is different about this circuit is following the first pentode with a triode.  This normally causes problems due to Miller capacitance, that is why the V76 et al. use pentodes for each stage.

The second thing I noticed is that they have switched in extra resistance into the first cathode circuit.  This is normally done with ac only, it is bad practice to switch DC IMO as it will generate switch noise.

The part that Abbey managed to lose, was about the optimal current for the EF86,  he said that 1.5mA was better whereas all manufacturers have found that around 0.6mA has less noise, Ian, I believe you have found this too?

From the magazine article that Rock kindly posted, it looks like  the S:N ratio was only 69dB.

DaveP

 

[abbey road d enfer]

The part that Abbey managed to lose,

Again, my apologies about this.  :-[

was about the optimal current for the EF86,  he said that 1.5mA was better whereas all manufacturers have found that around 0.6mA has less noise, Ian, I believe you have found this too?

That is not exactly what I wrote. I pointed that all the datasheets I know describe typical operation at 1-2 mA, and graphs tend to be centered on that. My concern was that it is almost impossible to predict the parameters of reduced current operation, and certainly not how partition is affected.
Analysing vintage gear is not necessarily enlightening; the guys at EMI knew their stuff, didn't they? They operated the EF86 at 1.35mA. OTOH IRT/TFK seemed to be more in line with the Lorenz designers, with 0.6-0.8mA. Maybe it's the German school against the English school.
There maybe other motivations than noise performance; it may have been a major concern for German broadcast who were deeply into classical. English engineers may have been more concerned with headroom and dynamic recuperation.

From the magazine article that Rock kindly posted, it looks like  the S:N ratio was only 69dB.

My understanding of the article was that the EIN was around -115dBu, which is not a particularly good performance; however, the reviewers suspect it's due to the tube's aging, which may or may not be true...owning an APx555 does not necessarily confer expertise.

 

[DaveP]

The part that Abbey managed to lose,
Again, my apologies about this.  :-[

No problem my friend, just teasing

Looking at the R47, it is very similar to the Lorenz, pentode followed by doubled up triode with DC feedback anode to anode or shunt as Ian prefers.

I am in the process of testing tube noise in my workshop at the moment.  This is not an easy job.  First thing I had to do was get an incandescent lamp to use for testing as the fluorescents caused too much noise.  Any tips will be appreciated!

DaveP

 

[Ericbazaar]

Hello,

i have two Lorenz V241 here for Service. Stunning sounding Preamps. I can make detailed pictures if needed.

Best

 

[emrr]

Thanks for posting.

In America, it was frequently referred to in the 40's-50's as plate to plate feedback.  Simple description of the connection.  You don't find it much. 

Maybe I am simplifying Dave's statement, but there are quite a few preamps with pentode front ends feeding triode output stages. I found in RCA BA-11 the resistances are near the allowable limit, and NFB flattens treble response, it can't run open. 

Bad practice switching DC: yes, but this was not a switch that was frequently moved like today. 

If the person testing didn't have a half dozen or dozen samples of burned in and recently used tubes to check for overall EIN, and also let the amps warm up sufficiently, there's no telling what EIN could be.  If the first stage plate and cathode resistors are carbon, it's almost a given they will have increased in noise contribution with age. 

 

[DaveP]

i have two Lorenz V241 here for Service. Stunning sounding Preamps. I can make detailed pictures if needed.

That would be very helpful Eric

Please post as many pix as you like.

DaveP