DHT Audio Shootout @ London

Here’s the summary of the shootout day we had yesterday at my place with  Tony Rees and Andy Evans

Test Music

For testing the pre-amps Andy suggested the following 2 tracks:

1.  Ravel Feria (Cleveland/Boulez)
2. Janacek The Cunning Little Vixen Act 3 Scene 7 “A vixen’s running through the woods” Prague Theatre/Gregor

The Ravel is orchestrated for a large orchestra and percussion, and listening was particularly for the timbre of strings, brass and woodwind and the clarity of the percussion. The Janacek is for several different voices and chorus, and listening was for the tone and character of the voices and how well they were differentiated from each other. In both cases the overall clarity, life and crispness was considered.

Andy Evans

Although I wasn’t familiarised enough with these 2 tracks myself, I found them to be an excellent choice and a reflection of Andy’s musical experience and knowledge as musician. 

Ravel’s Feria provides a very complex arrangement of instruments and full breath of dynamics between strings, reeds and brass instruments which allowed us to assess the impact on timbre, detail and distortion of each stage. On the other hand, Janacek’s track was a good test on voices and choruses. Different voice tones and channel separation. I’m not fond of this track myself, but it was good for testing purposes.

Equipment Overview

Since I moved house, I set up a temporary system which excluded the Starlight CD Player so source is restricted to either vinyl or hi-fi streaming (Tidal or from Mac) into an old O-DAC. The O-DAC was connected into preamp A and then into preamp B into the 4P1L PSE output stage. The output of the O-DAC was insufficient to feed a 2 stage setup so needed a preamp into a preamp/driver. Listening was done in 2 sessions.

My pre-amps all have gyrator loads and various biasing, while Andy’s all had plate chokes and filament bias. Russian 0.2uF Teflon coupling caps were used in all cases. Filament supplies were choke input and PSUs used tube rectification, chokes and all polypropylene caps.

My pre-amps all have gyrator loads and various biasing, while Andy’s all had plate chokes and filament bias. Russian 0.2uF Teflon coupling caps were used in all cases. Filament supplies were choke input and PSUs used tube rectification, chokes and all polypropylene caps.

Listening Sessions

Morning Session 1.

Preamp A – Andy’s 10Y, 126C x 2 as plate choke, filament bias

Preamp B – Ale’s preamps as shown below.

01A, SiC diode bias, gyrator

The first pre-amp we tried was my newly born 01a stage with Rev07 Gyrator board and cathode bias with an array of SiC diodes.

Since this was the first preamp there was nothing to compare it to, but it was clearly good, so was put aside to hear again

Andy Evans

UV201A, filament bias, source follower

This incarnation of the 01A preamp is based on the old brass-base, short pin, UV-201A. They are very old valves which cannot swing too many volts as linearly as the 01A or even the 10Y/VT-25. We found it a bit disappointing at the treble with the strings. Perhaps if we would have switched the order of preamps this may not have been that evident. I should look at this test again. 

Sounded bland though still clearly a DHT. Violins were rather papery and there was some edge in the sound. This is a very old pair of tubes and probably not quite the equivalent of later 01As.

Andy Evans

VT-25, filament bias

This is my favourite stage alongside the 01A. I found this pre-amp to provide the best detail in a DHT stage. I attribute this to the thoriated tungsten filament which has alongside the 01A. 

This was solid quality and well liked – smooth and with good timbre and detail.

Andy Evans

EML20A, filament bias

I was surprised with the stereo imaging and instrument separation that this valve introduced which was different to the rest. However, disappointed later as the timbre on instruments wasn’t as good as the other preamps.  I suspect that some of the underperformance of this stage was due to the fact it was acting more as driver (as second stage) than as a preamp. It has a low bias voltage (between 3 and 4V) which may have pushed the system into A2 operation during the louder sections of the test track and made brass, strings and cymbals to sound brighter and less enjoyable.

This promised to be interesting but was somewhat disappointing. It was bright-ish and lacked a distinctive timbre in the woodwind and voices, plus the percussion was less clear. Strings were OK. It was superficially impressive but soon started to reveal signs of overall weakness. More generic and less of that magic DHT sound than the 01A, 10Y or 2P29L.

Andy Evans

6E6P-DR, un-bypassed cathode bias 

This pre-amp is the best example in my opinion of what a DHT brings to the table: sound detail and timbre. The 6e6p-DR is probably the best IHT stage I’ve built so far and I highly recommend it as a driver. It doesn’t bring the complexity of the DHTs in terms of supplies, etc and provides a well rounded sound and is very pleasant to listen to. Its gain is good as well as its linearity. I think it has a better role as second/third stage in a phono preamp where DHTs may be hard to implement due to microphonic noise constraints.

This was very good for an IDHT. The overall sound was pleasant and easy to listen to and like. However, compared to the 10Y the strings lacked that DHT level of clarity and the timbre of acoustic instruments and voices was rather generic. It was not quite on the level of the EML20A but not far short and generally more pleasant on the ear. For an uncritical listener this could be a good choice, especially since it could drive a 300B or 2A3. But that particular DHT magic eluded it.

Andy Evans

2P29L, filament bias

This was probably my best discovery over the last years in terms of all rounded DHT. It has excellent detail and provides very good timbre all round on woodwinds, voices and percussion. It’s electrically easier to implement than any other DHT and has no microphonic noise whatsoever. The unbeatable advantage is its cost. On the flip side, it doesn’t have that magic on timbre and detail that a thoriated filament valve does, so isn’t my number one choice.

Good timbre to woodwinds and voices, though climaxes were just a touch grey and thick. The top end just lacked the ultimate clarity of the thoriated 01A and 10Y, but otherwise the sound had that DHT magic and was 90-95% that of a 10Y. Kind of a poor man’s 10Y that could be left on all day with no worries.

Andy Evans

01A, SiC diode bias, gyrator (Second time)

We decided to hook up again the first preamp which surprisingly was for all of us the best sounding. And we were surprised again as Andy clearly points out below:

Coming back to this preamp was quite a shock. It was pretty stunning and clearly the best of Ale’s builds. Everything was better – timbre, clarity, voices and overall likeability. This remained in the system as first choice and was then transferred to Preamp A

Andy Evans

Afternoon Session 2

The shootout turned to be quite labour intensive and exhausting for our brains. When you’re hooking up and swapping stages and different supplies you don’t want to make any mistakes. We’re dealing with High Voltage and a mistake may result in damaging one piece of equipment which we may regret for sure. 

After the break, it was nearly 2pm and we were hungry. I offered a quick barbecue which was turned down by my guests as it sounded too much work and everyone were keen to get back to the listening room. Instead, a pizza and a bottle of nice white wine was the chosen menu. 

Preamp A – Ale’s 01A with SiC cathode bias

Preamp B – Andy’s preamps

10Y with Hammond 126C x 2 in series as plate chokes

This was a good sound, similar to Ale’s 10Y preamp. Clear and satisfying smoothness, clarity and timbre.

Andy Evans

2P29L in filament bias with NP Acoustics large amorphous plate choke

180H Chokes / 40mA running at 14mA

This was another good sound, with the expected clarity and good timbre, about on a level with or slightly under the 10Y 126C.

Andy Evans

2P29L in filament bias with NP Acoustics small amorphous plate choke

136H Chokes / 15mA.

I found personally this preamp to be the best of the 2P29L choke loaded versions that Andy built. 

This was again a good sound, with the expected clarity and good timbre, a level slightly but audibly above the previous two preamps. This was the best of the plate chokes, in this case running at 12mA.

Andy Evans

2P29L with Hammond 126C x 2 in series as plate chokes.

Again a good DHT sound, on the same level or slightly below the larger NP plate choke but inferior to the smaller NP.

Andy Evans

Overall best sound

The combination of the 01A with SiC diode bias and the 2P29L with the smaller NP amorphous choke was the best of the day. We switched these two preamps in place A and B and results were generally similar.

Andy Evans

Due to my system limitations during the listening day, we didn’t test all stages in the same way. I’d have rather leave the 10Y/VT-25 as driver stage and swap the preamp A instead. In this way we could have assessed better the differences between the 01A (SiC version) and the 2P29L. 

Overall rank order:

1. Ale’s 01A with SiC diode bias was the clear winner
2. VT-25/10Y with either gyrator or plate choke, 2P29L with NP small choke
3. All the other 2P29L preamps
4. UV-201A, EML20A
5. 6E6P-DR

Conclusions and final thoughts

These were a good bunch of 10 preamps, with no weak ones. Differences were audible though in some cases quite subtle. The tubes used all had merits. The SIC diode bias shows the way forward, judging by the results we heard, and some quick conversions are being planned to some of the preamps! We preferred the thoriated tubes overall – both the 01A and the 10Y – for the added sparkle and clarity they gave. The 2P29L was not far behind. Many thanks to Ale for hosting and Tony Rees for critical listening and feedback.

Andy Evans

It was a great and exhausting day. We didn’t relax and sit down to listen music as we normally do, but the effort was worth every hour. 

I learnt yesterday that what you put in the cathode of an DHT is more relevant than as the load. Gyrator and chokes when selected and used properly at their best, sound very close. However, the difference is in the DHT and the bias arrangement. For years we followed the path of filament bias. Lately we settled on NOS wirewound Russian military resistors after some listening tests. It was evident then that the quality of the resistors made difference.  However, we avoided were we could cathode bias due to the requirement of an unwanted bypass capacitor. 

I did a lot of experiments lately with un-bypassed cathode bias (or cathode degeneration as you may prefer referring to). I found noise issues and bandwidth limitation due to CCS-based regulators with resistors above 100 ohms. This isn’t a problem with IHT. I will write about this another time as it’s long (but very interesting). This experimentation took me down the path of testing back the SiC diodes which I’ve used in the past successfully. The result was amazing and unexpected. Still have to try using them in filament bias in replacement to the filament resistors. They may sound as good as they do as cathode resistor replacements.

Here is the circuit I implemented. Instead of Rod Coleman’s regulators I used a pair of  voltage regulators I designed myself, but it works same way with Rod’s regs:

I want to thank Tony and Andy for coming along and for their contribution. We learned a lot as a group and is invaluable their contribution to continue refining the DHT journey.