417A-300B Plate Choke or Interstage Coupled?

[sgerus]

After reading 100’s of posts on the topic of input/drivers for a 300B I have come to the conclusion that there may be a better valve to drive a 300B than the 6sl7 or 6sn7.

So, with that being said, I’m thinking of replacing my 5691 input/driver stage of my 300B amp with a 417A.

Question: Which configuration do you recommend?

1. 417A interstage coupled to the 300B
2. 417A with a plate choke cap coupled to the 300B
3. 417A with a constant current source cap coupled to the 300B

Thank You

 

[i_should_coco]

417A coupled by your favourite method to a 46, transformer coupled to a 300B.

Just my 2p.

 

[kevinkr]

Depending on who made your 417A/5842 mu is between 43-50, unfortunately the 417A is not going to swing enough voltage to drive a 300B to full output directly. Here an IT is your only recourse and to do it a ratio of at least 2.25 seems indicated - higher values resulting in better linearity at the expense of HF extension. (Higher reflected source Z driving the 300B grid.. IMO anything much > 10K is going to hurt the HF extension with 1:2.25 it's already around 8.6K) The 417A/5842 is an almost ideal IT driver for an application like the 2A3 or 45 where the drive voltage requirements are more modest.

Consider a 6CK4 driving an IT, it will swing at least 400Vpp clean on a 300V supply with a 1:1 IT.. Mu of 10, so it will not be hard to drive..

 

[arnoldc]

Hi,

I have experience with #2 and #3.

My 417A-45 uses plate choke (custom made) on the driver, cap coupled to the 45.

My 417A-300B/10Y/50 uses diyAudio CCS on the plate at 10mA, driving any of the three output tubes I mentioned. The output tube has a grid choke (for the 50 actually). In 300B mode, I run it at a modest 6W.

I like them both.

ps.

My 45 amp uses WE 417A, and my 300B/10Y/50 amp uses Amperex gold pins 5842.

 

[Tubelab_com]

Depending on who made your 417A/5842 mu is between 43-50, unfortunately the 417A is not going to swing enough voltage to drive a 300B to full output directly.

That is news to me. I use a CCS loaded 5842 to drive a 300B (or 45 or 2A3). The later amplifiers (all of the TubelabSE boards) include a mosfet buffer to allow A2 and eliminate transient overloads. Early in the design phase I tried conventional resistor loading of the 5842, and I could not drive a 300B into clipping.
Schematic:

http://www.tubelab.com/TubelabSE.htm

I run the CCS directly off of the main power supply and set the plate voltage on the 5842 to about 175 volts. I can drive a 300B into clipping directly from my CD player on some CD's. On other CD's there is not enough gain to reach clipping without a line stage, but there is plenty of drive voltage available.

 

[kevinkr]

Hi George,
I'll admit to being a perhaps little more conservative in their use than you are.. I generally run them at 150V and below (125V typical) based on the specs I have and what I had to pay for the last few I bought (kinda expensive for such a little chunk of glass ) - at these voltages the drive versus distortion seemed a little marginal in the samples I had on hand. (Raytheon) Unfortunately details no longer to hand..

 

[sgerus]

Thank you all for the replies…

Last night (before reading these replied) I did a load line plot of the 5842.
Looks like best case is about 1V=27V based on using a 12mA ccs.

With a coupling cap or 1:1 interstage I would need 2.1V to drive the 300B at -61V
This might work in my case (running off my pre-amp output), but I would prefers an input sensitivity of around 1V in order to run the amp right off an mp3 player.

I guess the answer to my original question really depends on what input sensitivity you need.

Lesson Learned: Don’t trust every schematic you find in the net…. I asked on question based on a schematic using a 1:1 it driving a 300b at -74V

I’ll check into the 6ck4 (and the 6aq8 as in the electra-print directive reactance drive 300b)

In general, would you guys agree that moving to a high current tube is a step in the right direction?

 

[Tubelab_com]

Mine are all Raytheon 5842's. With 175 volts on the plate, the voltage swings to about 300 volts on the positive peaks. I have not seen an issue with this in about 20 amplifiers.

A few years ago I was given a truck load of "worthless TV tubes", and another truck load of military surplus tubes. Most were indeed junk, with about half physically broken. There was a box with about 100 Raytheon 5842's. I have been using them in amplifiers ever since. A CCS loaded 5842 is one of the most linear tubes that I have found. Since the audiophile community has discovered this tube the price has risen sharply, and my box is getting close to empty.

I have tried dozens of other tubes to find another tube with similar characteristics. A few have come close, but none have matched the 5842.

 

[kevinkr]

I'm totally envious, I have just six 5842 and while they are still cheaper than a 417A they ain't cheap..

 

[protos]

Hi tubelab,
Your powerdrive circuit is going around in my head for a year now but other projects have taken precedence..I tried the ccs loading and I liked it.But ccs loading is somewhat more acceptable from a purist point of view.I wonder how running the signal thru a mosfet follower impacts the sound compared to normal tube driving.OK I understand your point about most tubes not having enough current to drive A2, but what about the rest of the time?
Do you only listen to rock music/live guitar for evaluation?
By the way another question:when is a tube suitable for A2 drive?I guess certain grid parameters are involved.Has anyone tried 6c33 with A2 and is this ok?
An alternative to 417 and 5842 would be the 6c45p probably.I tried this as a (ccs loaded) driver for 6c33 but it just misses swinging enough voltage in practice to drive a grid of around 90vpk.

 

[sgerus]

Kevin, the 6CK4 looks like a power triode. Is that the right number?
Did you mean to use this tube as an input/driver?

Tubelab, just so I understand correctly.
Your set up the 5842 with a ccs at 175V -2.25V 20mA
That swings from 80V-300V
(300-80) / 2 /1.414 = 77V
77V/2.25=34.5V

So to drive a 300B
1.7v =61v
2.1v=74v
This looks better

 

[Tubelab_com]

OK I understand your point about most tubes not having enough current to drive A2, but what about the rest of the time? Do you only listen to rock music/live guitar for evaluation?

Most driver circuits do not have the capability of sourcing current into the grid of an output tube. Even if they did, there is a serious problem if the output tubes grid is coupled with a capacitor. A detailed explanation of this "blocking distortion" can be found here:

http://www.aikenamps.com/

Go to "Tech Info" then "Advanced" and read "What Is Blocking Distortion". Lots of other good info here also. Many of this "guitar amp info" applies to our HiFi tube amps.

The standard method for solving blocking distortion is usually a driver transformer, but good ones are expensive. Cathode followers with direct coupling to the output tube are also used, but this often falls short of great sound. I came up with the PowerDrive circuit as a low cost method for avoiding blocking and found some extra power as a side effect (I get 40 WPC in my 845 amp). For 99% of the time the amp stays in A1. When a transient comes along that drives the grid positive, a normal (A1 only) amp will clip. If it is cap coupled there will be a recovery time after the clip where it will distort. This can be from a few milliseconds to hundreds of milliseconds in extreme cases. An amp designed to handle the grid current transient will allow (in this case) an extra 2 db of headroom before clipping, and the recovery time (if clipping does occur) will be zero (or a few microseconds).

When I started using the mosfet buffer a few years ago. I conducted several listening tests with a wide variety of music using 300B tubes. These involved a switchable bypass on the follower for near instantaneous comparisons on the same music. ALL reviewers liked overall sound of the PowerDrive circuit. Its effects were not noticable on simple music (a single vocalist) without large transients. It was deemed very effective on highly transient music. Not just rock music either. Anything with large transients. During one of the early sessions I remember playing Dave Brubeck (take five) over and over again because one of the reviewers kept hearing things that he never noticed before. If there is a circuit on my web site in the "Tested and Proven Circuits" section, it has been bench tested and listened to by several people. There is some new stuff coming, but it won't get published until it has passed some serious testing and listening sessions.

My personal tastes do favor rock music (about 70% of my music collection) but I listen to just about everything except rap, country, and opera! I have heard quite a few DHT amplifiers that can play Norah Jones very well, but only a few that can do justice to her music, AND play Metallica at the edge of clipping without making a mess of it.

My last test session was chronicled here (mostly older Jazz music):

http://www.diyaudio.com/forums/showthread.php?s=&threadid=90655&highlight=

By the way another question:when is a tube suitable for A2 drive?I guess certain grid parameters are involved.

I have not found a reliable way to determine which tubes like A2 from the data sheets or plate curves. If curves are given for positive grid voltages, and the linearity looks good (same spacing between curves of equal grid voltage steps in the range of interest), then chances are good that the tube will like A2 and you will get a power output benefit from using A2. Curves for positive grid voltages are not often given for many tubes. In my experience the PowerDrive circuit will always improve transient response and eliminate overload recovery effects, but may not offer a power output increase. This seems to be the case when using tubes that operate near the limit of the cathodes (filaments) ability to emit electrons. I have found this to be the case with the 45 tube. At least with the tubes that I have, the cathode current will limit at 50 to 75 mA no matter how far positive you drive the grid. When the grid goes positive IT draws some of the cathodes current, and the plate current actually goes down! I have no experience with the 6C33. You may just have to test it and see what happens.

I have used the TubeCad SE amp cad simulator a lot. It accuretely predicted the 40 watts that I get from an 845, but tells me that I will get a power increase by running A2 on triode wired 6L6GC's. I don't see that power increase.

When I designed the PowerDrive I went through about a dozen different mosfets to find one that has a very low transfer capacitance that doesn't vary a lot with voltage. I tested the potential candidates in the circuit (for flat frequency response out to 300KHz and good 50KHz square waves). Then they were sound tested in the previously mentioned listening sessions. I chose the Toshiba 2SK2700. These have been used in at least 10 amplifiers that I have sold, and a bunch of TubelabSE boards assembled by others. I have received notice (after they were gone) that the part is no longer manufactured because it contained lead. I am now testing several potential replacement parts.

Tubelab, just so I understand correctly.
Your set up the 5842 with a ccs at 175V -2.25V 20mA
That swings from 80V-300V
(300-80) / 2 /1.414 = 77V
77V/2.25=34.5V

I set the current source to 12 mA (with a current meter connected in place of the 5842, plate to cathode). Then I set the plate voltage (with the 5842 installed) to 175 volts by adjusting the cathode resistor. My batch of 5842's seems to have a wide variation from tube to tube, so the cathode voltage will wind up anywhere from 1.75 to 2.5 volts. This unfortunately rules out LED bias without trying a lot of LEDs.

I follow your reasoning up to the 77 volts RMS out of the tube. I am not sure where you got the 2.25 volts from, and I assume that 34.5 is the expected voltage gain. I have measured a voltage gain from 35 to almost 40 depending on the tube. They vary, and mine are ALL Raytheons from the SAME date code, out of the same military bulk pack!

The 6CK4 is a power triode used as the vertical output tube in old TV sets. It can be used to drive an IT. Other similar candidates are the 6AH4, 1/2 of a 6BL7 or 6BX7, or a two stage driver using a 6EM7, 6EA7 or 6DN7.

 

[sgerus]

I am not sure where you got the 2.25 volts

If the cathode is at 2.25V, 77V(out)/2.25V( in -the swing from 2.25 to 4.5)= 35V
ie 1V in = 35V out

 

[arnoldc]

I have tried dozens of other tubes to find another tube with similar characteristics. A few have come close, but none have matched the 5842.

George, while I'm a fan of 5842, I was wondering if the 407A (2C51) would come close or better. Have you played with these yet?

I'm totally envious, I have just six 5842 and while they are still cheaper than a 417A they ain't cheap..

Kevin, what's the going rate for 5842 now? While my friend is not as lucky as George, he was able to buy 200pcs of Raytheon 5842 with a couple of Amperex 5842 gold pin Orange label and Green label at PHP200 each... about USD4.

 

[Tubelab_com]

I was wondering if the 407A (2C51) would come close or better. Have you played with these yet?

I have a zillion 2C51's or 5670's. I tested them a while back and decided that they needed further investigation.

I went through my warehouse and got several samples of all the tubes that I have in large quantities. I started testing them for noise, distortion, output voltage capabilities, gain and frequency response, with a resistive and CCS load. I put all of this information in a spread sheet, and sorted it in order of merit for several different applications. Unfortunately this spread sheet was lost when my old computer died last year. I remember the 2C51 being near the top for some uses, and I have been using them in LTP phase splitters, but I need to re test them for applications where large voltage swings are needed. The lowish plate dissipation ratings limits the ability to run my usual 10 - 15 mA of driver current, unless I parallel both sides.

When I was designing my SimpleSE board (one triode section driver, CCS loaded, no mosfet follower), I considered the 2C51 (and a few others), but I got better results from a 12AT7, primarilly because I could run higher current and it has more gain. It also has better worldwide availability (and I have a zillion of them too).

5842's currently go from $6 to $15 USD. There are vendors advertizing them for $6, but they never seem to have any in stock. Some purchasers of my boards have reported paying up to $25 USD each for WE 417A's

 

[dsavitsk]

5842's currently go from $6 to $15 USD. There are vendors advertizing them for $6, but they never seem to have any in stock. Some purchasers of my boards have reported paying up to $25 USD each for WE 417A's

Do you have any sources you can share? Tubesandmore.com has them for about $13, but most of the other usual suspects (tubedepot, tubestore, etc) are more like $25 and up for the raytheons and $70 for the WEs.

 

[arnoldc]

Hi George, thank you very much for the elaborate information! Much appreciated.

 

[Tubelab_com]

Do you have any sources you can share? Tubesandmore.com has them for about $13, but most of the other usual suspects (tubedepot, tubestore, etc) are more like $25 and up for the raytheons and $70 for the WEs.

I have never actually purchased any 5842's or 417's, since I still have a few 5842's left. I am relating the prices that I have been told in e-mail that I have received relating to the TubelabSE board. I checked my usual tube supplier, and he has none in stock. He did tell me that the prices have gone up considerably in the last year, so your numbers may be more realistic.

When I designed the TubelabSE board, 5842's were $5 on Ebay, and $6 from my usual supplier. That has changed, and I get e-mail every day asking if there is any other tube that will work in the board. I have been looking for a driver tube that works as well, and is generally available. When I find it, a redesign of the board will be required.

 

[cerrem]

Just to add in my 2 cents....
I have used the 6V6 as a driver valve for SET amps with good success.... I wire it as a follower....I sometimes use feedback from the output stage and apply it to the cathode tail to linearize the drive...bootstrapping method...

Chris

 

[Tubelab_com]

The 5842 (CCS loaded) has enough gain to go from CD player (or linestage) to 300B grid. I have found a few tubes that may do the job, but they are not cheap either, If I wanted a two stage driver circuit,the first 2 stages would be in a 6EA7.