As always opinions are very personal in this subject, however there is a shared view out there that DHT sound is very unique. I do like DHT pre-amps and have tried all sorts of configurations and many valves (e.g. 46, 10Y, 01a, 30, 30sp, 71a, CX112, 26 and 4P1L). After testing many of them I enjoyed most out of the 4P1L (Siberian pre-amp), CX301a and 26. The 26 is very special to me so decided to improve the previous version incarnated around the LL1660/8mA.
This time I decided to use the LL2745 provided by Thomas Mayer. Again, filament bias and Rod Coleman boards were the foundation of this design. I replaced my Salas shunt regulator for a simple stage of Russian glow tubes fed by a CCS (a la Gary Pimm). The test results were very promising, so this became the main HT regulation for this preamp. I added the Morgan Jones’ famous Kelvin Caps as shown below:
R3 has to be adjusted on test to ensure the stack of two glow tubes draw about 10mA and the remaining is supplied to the 26 amp (about 5mA). Glow tubes are bypassed by russian PIO C3 and R5 is fitted to ensure both glow tubes strike as expected.
I’m using balanced connectors now despite the input is unbalanced. A special cable is required to fit this input. Rf is classic well known dale wirewound stackable 10W resistors. I’m currently using a pair only as filament starvation reduces the power dissipation requirements.
The output is balanced and floating. It can be wired to provide unbalanced if needed. I’m planning to build in the future a balanced input PP 6C4C amplifier so wanted to have the preamp ready for this amp.
The 26 is running slightly starved. I found an operating point that minimises distortion, microphony. Filaments glow will be very dim here.
The output transformer is wired in 5.6:2 mode to maximise the gain of this stage. This comes out paying the price of additional output impedance. In this case I’m not driving an TVC or long cables so is not an issue. Otherwise, the OT should be rewired to 5.6:1 which will give a lower output impedance around 300Ω, but lower gain.
The HT and LT raw supplies are unchanged, and are same ones from original design. You can obviously improve these, but I found the pre-amp to sound very nice so will keep them as they are:
Filaments are very quiet using Rod’s DC regulators. The noise at the filament is about 180μV. Originally I added an RC network at the output of the glow tubes which is not ideal as increases the output impedance.
After fitting the Salas SSHV2 regulator the following measurements were taken:
- Operating point:Va=145V, Ia=4.5mA.
- The supply ripple measured at the anode was -85dBu (50μV)
- Filament noise/output noise with filaments only: 440μV/180μV
- THD: 0.03% @ Vo=3Vrms (9dB gain)
- THD:0.01% ” Vo=1Vrms
- HT ripple: -80dBu (100μV)
How does it sound?
Well, what a tricky question. The typical answer from any consultant applies here: “it depends”. You can’t compare this component of the sound reproduction chain in isolation. Yet, I can provide my view based on other preamps used with the same system. The 26 pre-amp sound is very clean and warm. Has a good bass, overall tone and nice treble as well. I’d say I prefer slightly the sound of the CX301a preamp with gyrator load. Still need to listen to this pre-amp a bit more.
As usual my big thanks to Andy Evans, Rod Coleman, Thomas Mayer, DHTRob, Vyacheslav, Dmitry Nizh and Paul Leclerq
26 DHT Gen 2 builds
Rui Lourenço built a fantastic version of this preamp. If you’re interested on this, check it out here.