The EMP-351 - an Ampex 351 inspired dual microphone preamp scratch build

I realized today that it's been nearly a year since this blog has been updated, but that doesn't mean that Wombat Amps hasn't been busy!  On the contrary, we've now launched an improved web site, I've established a formal Wombat Amps workspace in a building that also houses dozens of bands -- as a consequence I've been doing a fair amount of repair work -- more than I had really intended actually.  Nevertheless, I have been working on some interesting projects, including a cool custom amp that was recently completed and a solid state equalizer.. But details of these just haven't made it to the blog for want of time.

However, I'm working on a new microphone preamplifier project that I'm really excited about, and I hope to document a fairly complete build log here.  This mic preamp was commissioned by Andrew Everding, perhaps best known as the keyboard player for the band Thursday. However, with Thursday currently on hiatus, Andrew has moved on to audio engineering, and is currently located in Auckland, New Zealand, where is an audio engineer working directly with Neil Finn (of Crowded House fame) at Neil's well-known Roundhead Studios.
 
In any case, Andrew is a big fan of tube microphone preamps based on modification of old Ampex reel-to-reel players.  As a tube aficionado, this makes perfect sense to me - in a world now flooded with sterile-sounding ultra-low distortion preamp choices, more and more musicians are realizing that a bit of distortion, of the right kind, can be just the ticket for producing a warm, full sound.

One of the preferred platforms for the mod community has been the amplifier from Ampex 351 reel-to-reel player, as shown above.  In fact, there's a fair amount of documentation online (especially at the Electric Audio forum, courtesy of Greg Norman), describing how to mod one of these units into a preamp.  So why not just mod one of these, and be done?  Well, it turns out that there are a number of good reasons.  First, the Ampex 351 is now highly sought after, and prices for these units have skyrocketed. You can easily pay $1300 or more for one on fleaBay, and what you get will be in uncertain condition. And this is before the mod process, which turns out to be not at all straightforward.  These were constructed on now crufty old first-generation PCBs that are prone to traces lifting from accumulated moisture, and are not particularly amenable to modification.  Also, there is a ton of other circuitry in there that gets in the way and is unneeded after the mod, you're for sure going to have to replace all the power supply capacitors anyway, etc., etc..  Even after you go through all the trouble, you're still going to be left with just a single preamp, not a dual. So you'd need to do this all twice to get a dual setup!  And this single would not really be optimised in terms of layout, ground scheme, noise, etc.  It also wouldn't provide modern amenities such as phantom power and fully regulated DC tube heaters, at least not without having to do a bunch of extra work.  But aside from all that, the Ampex is still a pretty damn good reel-to-reel player, and I just hate seeing beautiful old gear being canabilized in that way.   So the decision was made early on in the process to scratch-build from new components a dual preamp that stays true to its heritage but incorporates just the components necessary for a preamp.  In addition, it should incorporate features such as phantom power and regulated DC tube heater supply, and do it all in an attractive 3U rack-mountable package.  Here's the design brief:

• Dual microphone preamp inspired by the Ampex 351 circuit;
• Each section to combine elements of the Record and Repro boards of the original in a manner similar to the Electric Audio preamp mod;
• Tube complement of each section will be 2 x 12AX7 and 1 x 12AU7;
• Tubes provided will be New Old Stock (NOS) selected for low noise, gain, and balance;
• Custom wound transformers audio transformers from Sowter UK;
• Custom wound power transformer from Heyboer USA;
• 3U rack-mountable enclosure with laser-engraved faceplate;
• Black anodized aluminum contrasting with large high-quality brushed aluminum control knobs;
• Regulated 12.6VDC filament heat for all tubes;
• Tube rectifier replaced with silicon rectifier;
• Switchable between 120V/60Hz and 230Hz/50Hz power;
• Regulated 48VDC phantom power;
• Switchable -20 dB input attenuation on the XLR mic input;
• A separate 1/4" stereo jack line level input with fixed -30 dB attenuation;
• A front panel Mic/Line selector switch;
• A mono 1/4" instrument input directly to the input tube grids;
• Phase switch;
• Variable output attenuation using a  Bourns 600Ω T-Pad Attenuator;
• Bypass switch for disengaging output attenuation;
• Primary gain control and gain trim adjustment;
• Ruggedized turret board construction for main amplifier boards;
• Custom hand-etched PCB for phantom and filament power supply;
• The unit will be equipped with an internal EMI/RF line power filter;
•  Power supply features 105deg-rated low Z electrolytic capacitors for heat resistance and longevity;
• Signal capacitors are high-quality film types in the nF range, and silver mica in the pF range;
• Audio path wiring is aircraft-grade solid core 20AWG with PTFE insulation, shielded where necessary;
• Controls for gain and gain trim will utilize sealed 2W milspec potentiometers;
• 750V rated 2W metal film resistors will be used for low noise and durability;
• Where accurate balance is required for best CMRR, 0.1% precision resistors will be used;
• The ground scheme will be optimised for low noise, and will adhere to the AES48 standard to avoid the "Pin 1 problem";

We'll discuss the circuit itself in another post, coming soon! Please follow along as this project comes together.

Thanksgiving update on the EQP-1A project

Now that the Grail amp has made its way to Switzerland, it's time to tidy up a couple of loose-thread projects.  First up is the final completion of the EQP-1A tube equalizer project,  intended as a companion piece to the Wombat Amps version of the LA-2A compressor I constructed for Black Iris Music. 

The main boards for this one have been complete for awhile, as has been the laser engraved front panel.  So the main job has been getting all the hardware mounted, and beginning the process of harnessing everything together.   Like the LA-2A, this project will also feature DC filament heaters, which, in conjunction with metal film resistors and a modern layout & ground scheme, will ensure virtually noiseless operation - or at least as noiseless as is achievable with tubes. 

Yesterday I had a chance to put together a DC heater circuit, based on the NJM2396F63 low-dropout regulator.  It's an interesting part, because it is designed to provide 6.3VDC -- and there's only one reason why anyone would want that oddball voltage... it's the standard for tube filaments. I'm not aware of any other application that require this, so I'm surprised there's still enough of a market niche to make these worthwhile to produce. 

Anyway, here's where things stand:

 

 Completed back panel 

Note that the power transformer is mounted internally, as are the tubes.  Visible are the Sowter UK input and output transformers.  The input transformer is a 3603e 600:600 ohm isolating transformer.  The output transformer, the 9530e, is Sowter's version of the literally impossible-to-find Peerless S-217-D found in the original.  It's a highly unusual transformer, as it includes an odd centre-tapped tertiary winding that is used to provide about 4% NFB to the input valve, as is visible in this schematic.

Back Panel insides

Here's a view of the back of the inside of the back panel.  I'm using the Hammond 270CAX as the power transformer.  The current plan is to use the 5V and 6.3V windings of this transformer in series in order to feed the DC regulator.  Note that the 6X4 rectifier tube is slightly unusual in that it employs 6.3VDC for its filament, rather than the common 5V -- all three tubes together will require 1.2A at 6.3VDC.  By my calculations (and accounting for power factor) this will push the VA rating of the 5V winding a bit.  However, the rest of the transformer will be pretty much loafing - the current demand for the high voltage part is very small indeed.  I therefore don't anticipate any difficulties, but in the unlikely circumstance that the transformer does get a bit too warm, I have a couple of fallback strategies... I left enough room internally for a separate filament transformer, as I used in the LA-2A.  Alternatively, it turns out that there is a 12.6V version of the 6X4 rectifier (called, you guessed it, the 12X4) -- just changing to a 12.6V heater instead of the original 6.3V would halve the current demand, and would be easier on the LDO regulator as well.

Mounted main boards

Finally, I got the main boards mounted in place, and the majority of the front panel controls wired up.  I made one small modification the larger main board... you'll notice a silver can visible on the right hand side.  This is the Sowter 9330e interstage transformer, which replaces the original Triad HS-29.  There was just enough room to mount it here, and the position turns out to work well with my layout.  Finally, the silver can (actually mu-metal for best noise rejection) on the filter board with all of the orange-drop capacitors is the Sowter 9325e inductor required for the EQ.  This build sure required sourcing a bunch of very specific oddball parts, so I'm very grateful that Sowter is producing these extremely high quality reproductions -- otherwise a faithful build of the EQP-1A would be next to impossible.


It won't be long now! Look for a wrap-up on this one soon!









 

Coming soon - the Wombat Amps program equalizer 1A

It's on hold temporarily while I work on the Grail amp commission, but I've also been working on my next studio gear project --  a clone of the revered Pultec EQP-1A tube EQ unit. It's basically a nice passive filter circuit with a balanced tube recovery section.

As is usual, I've gone my own way with layout and ground scheme for this rack-mountable unit.   As I did for the Black Iris LA-2A project, I'll be converting the heaters to DC for exceptional noise performance.

In this case, I've made the decision to mount the three tubes internally.

I'll post layout images later, but in the meantime, here's some pictures of the already-wired main board and filter board.  Notice that for the filter board, I've chosen to mount the Sowter inductor directly to the board.   I'll also be using Sowter UK iron for the input, interstage, and output transformers.

Filter Board and Inductor:

 

Main board, front:

 

Main board, backside: 

Transformers for the LA-2A

Thing continue to develop on the LA-2A component acquisition front. The LA-2A has three transformers.

The easy one was the power transformer.  The one in the LA-2A is pretty much a bog standard PT -  the schematic specifies a 250V-0-250V  (500V CT) delivering 40mA, and 6.3V CT.   The center tap on the 6.3V isn't even really necessary, since I plan on using DC elevation of the heaters to reduce hum, and will be implementing an artificial centre tap in any case.  Frankly, there's not much mojo to be had in the PT - this is not like some guitar amps where the voltage sag of a poorly regulated transformer actually contributes to the sound.   So there appear to be lots of good alternatives.   Running down the list of usual suspects, it looked to me like the Hammond 269JX would do the job perfectly.

However, there are two audio transformers, and these almost certainly will make a noticeable contribution to the tone.  I did a LOT of thinking about these choices, only a few of which I describe here.

The originals at different times in the production run had two different input transformers.  The first was the UTC HA-100X, which has a fairly complex series of primary windings, which you can jumper in a variety of ways to get different impedances... 50 Ohm, 150 Ohm, 250 Ohm 333 Ohm and 500 (600) Ohm were all possible, although pretty much everyone, I assume, just wired it for a standard 600 Ohms. The secondary impedance is 60K. This transformer has a good reputation, but for all intents and purposes it's now unobtanium.  The other input transformer in the originals, and which was evidently more widely used, was the UTC A-10. Overall, its specs look fairly close to the HA-100X, except that it had secondary wound for 50K ohms. Authentic copies of this transformer in good condition seem to be pretty damn hard to source as well.

So, the hunt was on for a suitable boutique-grade modern production alternative.  The first option was to use the Jensen transformers  JT-11-P-1.  Although it is a 10K-10K bridging transformer, rather than the 600R - 50K of the original, many people have reported good results using it as an input transformer in LA-2As.  Jensen also has a simple schematic showing how to adapt the LA-2A circuit for this transformer.  On the one hand, it has a +20 dBu max input signal rating at 20Hz instead of the +15 or +16 dBu of the originals, so would have more headroom... probably a good thing now that signal levels tend to bigger these days than they used to be.  On the other hand, of the options this is the one that seems most obviously different from the originals. For some applications, it should work fine, but I decided to give it a pass.  For one thing, the 1:1 ratio would make it harder to use the LA-2A as a mic preamp with modest gain  -- which you should be able to do just fine by just pulling out the T4B optical unit! Also, I've heard reports that these units can sound their best when being pushed fairly hard.  And, er,  I also gave them a pass because the folks at Jensen couldn't be arsed to respond to my emails, and also didn't pick up the phone.

Next up  was the Cinemag CMQEE-3440A.  While the impedances look right, this was rejected because it looks to be intended as a mic input device, and not for line levels... I'd worry that it wouldn't be able to handle the signal levels without pooching out.   Someone suggested using the Cinemag CM-9589 10K:600 output transformer by wiring it in reverse, but that seems just too weird.

Another possible option was to have a company like Edcor wind a transformer to my specifications. On the plus side, they have good prices, and the guitar amp guys I know that have used them have been very pleased indeed.  But on the downside, I'm less familiar with their reputation for producing the high-bandwidth pro-audio stuff.  The real kicker though, was that because they custom wind every order, and usually have a backlog, it can sometimes take 6 weeks or more before your transformer is shipped.

Finally, I came across the Sowter transformers 1009.  This one looks to be an accurate reproduction of the original UTC A-10, with a +15 dBu input signal.  Moreover, there were really good reviews for LA-2As done with this transformer, and it was recommended in several places.  The only drawback is that Sowter is in the UK, and that delivery to the US is naturally expensive and relatively slow.  But at the end of the day, though, I decided to suck up the transatlantic shipping, and went with this.   This was also influenced by the fact that Sowter also turns out to make the Sowter 1010,  which appears to be an excellent quality 15K: 600 Ohm output transformer closely comparable to the UTC A-24.  As you might have guessed, the A-24 was the output transformer found in the original.   If anything, the 1010 has superior bandwidth.  Anyway, I plopped down my money and placed the order with Sowter.

Of course, no sooner did I do this than an original A-24 came up on Ebay...  It wasn't exactly cheap, but it wasn't outrageous either, so even though I'd just committed to buying the Sowter, I just couldn't pass up the opportunity to do a real-world A/B comparison with the original. In my experience, output transformers are where a lot of the magic happens. After dithering for a bit, I couldn't resist and ended up hitting "Buy Now" - an A-24 is presumably winging its way towards me now.   The plan is to let the customer do their own listening comparison, and let them choose whichever of the two transformers sounds best to them.   Maybe we could even do a blind comparison?  Anyway, it should be an interesting experiment....