A full write up of this preamp design and build can be found here
Tag: THD
4-65a SE Amp: refining the 46 driver
I did some tests today and looked at minimising distortion of this 46 driver in filament bias and found that Va=230V (instead of 184V) to provide best performance:
Filament bias resistor array is now laid out horizontally to improve the dissipation of heat.
Here is the performance (0.05% at 17Vrms) at maximum drive input from my audio test set:
4-65a SE Amp: testing the 46 driver
46 driver breadboarded. The mu-follower gyrator, the filament bias resistor array and the nice teflon UX5 socket from Jakeband. The filament bias resistor array is formed by 4 10Ω 20W dale wirewound resistors. These get very hot so probably need to think an alternative layout or further resistors in parallel:
The performance is very good. I just picked up a random 46 from my stock and biased it at 204V (which is the operating point in my design) achieving less than 0.05% at 10Vrms. Need to re-run this test to see how will perform at 70Vrms:
Using my spice model created from a good 46 valve, THD should be around 0.15% at 200Vpp with a 100K load and performance is great to loads down to 100 ohms. Clearly the load in A2 will change from high impedance to some kΩ so this driver should maintain outstanding linearity all the way through:
4P1L PSE tests
Testing parallel DHTs
After listening to a great incarnation of the 4P1L PSE in filament bias output stage from Andy Evans, I decided to have a look at the impact of unmatched pairs of triodes from a distortion point of view. Main reason was that when listening to Andy’s amplifier I noticed a bit of an uncomfortable treble with some strings. Perhaps the increase of odd harmonics, but wanted at least to see what was all about.
4P1L are very easy to match. you can easily get a pair with equal mu. Just randomly I picked from my collection a pair of valves with a difference of 0.5 in mu.:
THD is about 0.03% mainly driven by H2. It happened that one 4P1L from the pair had 0.02% where the other had nearly 0.04% distortion. The difference between H3 and H2 is about 8dB.
Then looked at a more closely matched pair (0.03 mu difference). The distortion wasn’t surprisingly different:
Again, nearly 0.03% and difference between H2 and H3 is down to 7.5dB.
Looking at the individual performance of the 4P1L, now biased at 30mA and similar anode voltage, we can see that despite having a lower THD, the difference between harmonics is just 5dB. This is the THD of the other 4P1L from the pair:
Well, how rthis compares to a 2a3/6C4C? The latter valves are two triodes physically connected in parallel inside the same envelope. So, no matching can be done:
The previous was a low distortion 6C4C I have. Distortion is higher than 4P1L PSE, but not that much. H3 – H2 difference is about 12dB.
My early thoughts:
- 4P1L are very easy to match
- 4P1L PSE performs really well. Distortion of the pair is lower than a 6C4C performing at same level.
- H3 component is higher in PSE and this could be the reason why is more noticeable when listening to strings – as I proved in practice.
THD, Gm and μ tester
As part of my repairing of the curve tester, I had to do some changes to the transconductance (Gm) meter section. Currently I’m leveraging most of the curve tester to also measure Gm, μ and distortion (THD). Albeit the latter is rarely used as I prefer an external equivalent CCS circuit that is not inside the tester as the output signal comes out cleaner. The curve tester provides all sockets, HT power adaptor, meters and bias supply.
This is my latest circuit:
The additional protection to fuse (F1) is the diode D2 which can protect the LCD panel meter A1 in case of an unexpected anode short. P1 and R4 were chosen to allow a precise setting of the anode current at low levels and some protection to the CCS when P1 is set to zero. M1 is bolted to chassis and is carrying all the effort when providing current at lower anode voltages. M2 on the other extent can be a TO-92 type. R1 was added to allow a bleeding path to C2 when not measuring transconductance. The bias section is a simple adaptation of Merlin Blencowe’s “Power Supplies for Tube Amplifiers”, which I suggest you take a look at as Merlin covers very well the most common valve bias circuits
With this circuit I can measure very accurately transconductance at any desired point. I highly recommend you Alan Douglas’ “Tube Testers and Classic Electronic Test Gear”, which has a lot of details around how classic valve testers work, challenges around Gm measurement and obviously some good ideas and suggestions for calibrating and measuring Gm correctly.
Obviously adding an amplifier to the Gm tester section could improve the accuracy of low transconductance valves. But that would be for another time!
Listening to the 26 preamp
Now that I have a very quiet preamp indeed, I’m a happy bunny. Changed today the pair of White Westinghouse 26 ST NOS that I’ve been using for over a month with a pair of RCA CX-226 and have to say that I enjoyed a far more sweeter and rounded tone out of this preamp.
Bass is still deep and treble is clear with the sweetness of the globe valve fingerprinted in the tone as you would expect. I will run this setup for a month and report results…
You can still see the clips used to provide earth return at the output to avoid the ground loop that I unnecessary created. God knows how may tweaks I did to improve the HT rail ripple before I realised that it was a simple ground loop.
26 DHT preamp with LL2745 OPT
Today I managed to finally test the Lundahl LL2745/8mA specially designed by Lundahl for Thomas Mayer for a 26/01a preamp / line stage.
The circuit tested is here.
Initially did some tests with fixed bias, normal DC heater supply from my workbench and HT from a passive capacitor multiplier also available in my workbench. Circuit breadboarded has long cables and we shouldn’t expect good 50/100Hz noise levels as have many transformers and things around 🙂
Test 1: fixed bias
- Vgk=-6,8V
- Ia=6mA, Va= 119V
- THD=0.033% @ Vo=1.4Vrms
;
Test 2: fixed bias with Rod Coleman regulator
- Vgk=-6,8V
- Ia=6.5mA, Va= 119V
- THD=0.031% @ Vo=1.4Vrms
- Vf=1.4V (minor starvation)
Test 3: Filament bias with Rod Coleman regulator
- Vgk=-6,3V
- Ia=6.5mA, Va= 119V
- THD=0.031% @ Vo=1.4Vrms
- Rfilament=5Ω, Vf=1.4V (minor starvation)
As we shown earlier in other tests, filament starvation reduces THD slightly as expected. The OPT performs really well. Probably will look at starving a bit more filaments whilst doing a listening test
Will now proceed to rebuild the 26 DHT preamp with LL1660 with this circuit 🙂
126C OPT
After testing 4P1L with the 126C OPT, Paul Leclercq suggested a different operating point to see if distortion above 8Vrms on the OPT version was due to grid current. So firstly did a quick test with the CCS and found that -15V and 170V was a good operating point for 15mA. The 4P1L performs brilliantly with CCS and THD is only 0.05% when driven to 10Vrms.
Unfortunately distortion with 126C grows exponentially above 9Vrms. OPT is tested floating and not sure if is the Pete Millett interface a limitation with it.
EDIT
Andy Evans suggested that secondary green should be negative so phase is 180. Aparently there is a thread meantioning the distortion introduced by the way I was using the OPT.
Anyway, problem sorted! Look at now how nicely the 126C can perform at 10Vrms:
4P1L with 126C OT load stage
Andy Evans built a pre-amp with the 4P1L and was delighted with the sound of it albeit the 4P1L was running below its optimal operating point: 15mA given the limitations of the 126C interstage transformer.
I went to my workshop to test this configuration and looked at biasing 4P1L with fixed bias and driving it with 1Vrms or more to see what the results were:
So here is the first test at Vg=-4V, Va=74V and Ia=15mA
(all tests were done with the 100k input impedance of the Pete Millett Sound Card interface as the secondary load of the OT. 4P1L had both filaments in parallel and If=600mA)
You can see a richer harmonic profile with the OT and distortion is around 0.13% when driven with a 1Vrms providing an expected Vo close to mu (Vo=8Vrms)
The distortion gets very high when output voltage is higher than 9Vrms:
Now if we bias the valve at a more convenient operating point:
We get a slight improvement in THD down to 0.11%. However the distortion above 9Vrms is still high:
So what if we compare the performance of the OT against the CCS?
As we can see from above the distortion is halved. Now if we look at how well this valve could perform if biased in a better operating point, we can see that distortion can be reduce down to 0.03%
Minimum distortion from a CCS (or gyrator) doesn’t mean that it will sound better. Clearly the OT doubles the THD of the CCS equivalent circuit. Gain here is nearly same on both as OT is in 1:1. Only way of judging both is to do a listening test….
(which is what I’m planning to do next)
4-65a SE driver
46 DHT in triode mode as a driver
Continuing with the design of the 4-65A SE amplifier based on M. Koster design, I’m in the process of tweaking the 46 driver to optimise the operating point and provide maximum distortion to drive the demanding 4-65a. Here is the circuit I’m currently working on.
The current 46 driver will be biased at around 25-30mA using filament bias, so Vgk will be around -16 to -17.5V using a 10Ω filament bias resistor array. This will set the 46 at around 185-210V which will give sufficient headroom (i.e. need about 200Vpp max) to drive the 4-65a.
So today I look at varying slightly both anode currents and Vgk to see impact on THD.
So minimum THD is around -17V and Ia=30mA. So if setting the Rod Coleman filament regulator to ensure that Vgk=-17V and the anode gyrator to set anode voltage to ensure Ia is close to bias current would provide the minimum distortion (which is 0.04% in this example). Pa is close to 7W, but looking at the datasheet we can see that maximum Pa is 10W (as the latter 45 version).
So next I need to build a prototype of this driver with filament regulator and gyrator load.