Saturday, 24 March 2018

EL84 Amp II: Progress Post 6 – Listening Tests

With electronic construction complete the amplifier has been moved from the workbench to the listening room to get some initial listening impressions. 

The speakers it's running into are floorstanding KEF C95s from around 1990. Efficiency is quoted at 90dB. These are sealed bandpass.

The amp has 3 line inputs and one phono. So the source used for listening tests was the Chromecast Audio and my turntable with its Audio-Technica AT11 MM cartridge.

First up - the line input. An easy test. Passed with flying colours. The measurements on the workbench were showing around 15 watts power into the 8Ohm dummy load before any sign of clipping, and the THD was reading less than 1% rising as the power level approached maximum. So expected behaviour, in other words.

The EL84s are biased to about 8 watts dissipation, which at the B+ of 365V means 22mA. The cathode shunt I am using is 10Ohm so that's a reading of 0.22V at each cathode test point.


EL84s with bias adjusters

The topology of this amp is that the first thing the line-level input signals encounter is the Cathode Follower which forms the first stage of the tone control. This means the input impedance is around half a meg. So a very easy load to drive. Most amplifiers route the input signal to the top of the volume control pot, so the input impedance is effectively the track resistance of that pot, usually 50 or 100K.



Hum levels are acceptable. There is a barely detectable hum if pressing the ear to the front of the speaker cabinet when no music is playing. However since I don't know anyone who would use the amplifier in this manner, I'm not going to allow myself to get too worried about it.

Sound quality assessed subjectively is clean, pleasantly detailed, no trace of any distortion or harshness in the treble, and with plenty of power to the bass. I would be happy having this sound quality as my daily driver.

The tone control behaved exactly as the previous build of this circuit, this is the second time I've built this one.

The phono stage however was not such a stellar performer.

The only problem with this one is an unacceptable level of hum. This is normal with phono stages and simply indicates a signal grounding problem somewhere.

Other than the hum, the RIAA stage sounded well. So I am confident some simple experimentation with the earthing will resolve the hum.

I tested the RIAA compliance on the workbench and found it to be within 1dB from 45Hz to 15KHz. This is acceptable.



One aspect of this build I am less happy with is the power transformer.

This is a Hammond 370FX and it's loaded to its rate current. Around 140mA DC load on the B+ and around 5A load on the 6.3v

However this transformer runs hot. After 30 mins run time it's noticeably warm and after 60 mins you can touch it but if you wrap your hand around it, you'll want to remove it after 3 or 4 seconds.

I found a few discussions on DIYAudio commenting about the heat output of these transformers, and to be honest it makes me reluctant to use them again, either that or I'll be sure to apply a generous de-rating margin next time.

Cosmetically I still have some work to do





The person who I am building the amplifier for has indicated his dismay at the power switch. Something with a chrome toggle is preferred. Finding something that will a) fit without fouling the adjacent EL84 socket and b) fit the existing 20mm hole, is going to be a challenge.

Secondly, on the right hand side, the hole in the front is for the remote control sensor. This needs to be permanently fixed in place internally and a piece of translucent white acrylic press-fitted into the hole.

Electronically I need also to arrange a mute circuit to avoid switching spikes going through to the speakers when toggling the speakers/headphones switch, and to avoid the open-circuit hum going through to the speakers when the switch is set to headphone (thus disconnecting the input to the preamp)

First order of business though: cure the ground-loop hum on the phono stage...

Saturday, 3 March 2018

EL84 Amp II: Progress Post 5 – Confessions!

It's been a little while since I posted an update on this project, and there's been a lot of progress, as well as one or two hiccups.

Thought I'd put up a few photos today since I've been taking plenty.

First up - I've had a few requests from people (offline, as well as on) for some photos of the PCB fabrication process. Since I took photos during the etching process of the power supply PCB, I submit for the admiration of sev'ral viewers [anyone get the obscure reference?] the PCB etching process in stages...


Etching is just starting



Copper exposed to the UV light is being slowly etched away


A few more minutes and it's nearly done


Finished! Ready for washing and drilling


The completed Power Supply board was already shown in the previous post, so these photos are a trip back in time *by popular request*

I did make up another little board though, by necessity. The motorised remote control volume control board I bought from AliExpress didn't work, so I had to buy another one, only this was a different type, and of course it needed a different voltage, that I didn't have to hand. 

I needed 9V DC for that board, which I also needed for the signal relay which switches the signal to the headphone stage. So necessity being the mother of invention, this was the result


5vac in - 9V DC out. 

A voltage doubler and regulator board. A couple of diodes, some capacitors, and a 7809 regulator. Plus a switch (on the other side) and on a board that's smaller than a SD card. 

So – on with the amplifier. 

First order of business was to get the top panel of the chassis ready. This involved a lot of measuring and drilling - the mounting holes for the boards and transformers, then the chassis punch for the valve sockets. A lot of swarf ended up on the floor during this process.



After getting the top panel ready, it needed to go to the laser etching workshop before I could do anything with it. This is to get the identifiers for the valves etched on - this design uses four different types, so it's important to know which type goes where!

Once that was back, it was time to begin assembly. Mounting up the transformers and sockets to the top, and circuit boards underneath. A delightful jigsaw puzzle, but everything fit together nicely and it was not necessary to utter any curses.


Transformers and output valve sockets in place



Starting to assemble the business end



All boards in place, ready for wiring up


Next the back panel needed drilling - this design will have four sets of inputs – three line-level and one phono, with the necessary separate earthing point. Also the speaker terminals will expose 4Ohm and 8Ohm outputs, and of course the standard IEC Mains connector.

Everything was mounted onto the back panel, just to make sure it all looked OK and didn't foul anything inside the case when in place (it didn't, so again, no cursing necessary!)

Then it was a case of removing all the terminals and sending the naked panel off to the laser engraver to get the descriptions and other vital pieces of information added to it.


The back panel, before laser engraving


Then, just because it would be remiss not to, it was time for a photo session


Front panel legend - this is a valid design technique and don't let anyone tell you otherwise!






Showing the bias adjusters and test points for the output valves (same design as the last EL84 amp I made)


So at the top of this post I mentioned one or two hiccups. This firmly comes under the category of "learning from mistakes". Those who are more experienced at this may choose to laugh at my misery if they are of a vindictive nature, or sympathise if they are more empathetic... but I screwed up the low-voltage side of this amplifier rather badly and it's going to need a rather ugly (and obvious) rescue.

I will disclose my thinking and why it didn't work here, in the hopes it might help someone.

Warning: There are no more pictures, and it gets a bit technical from here on.

So going by the previous pictures you will see there is a lot of glass here – 13 valves to be exact. This is because of the configuration - a RIAA stage, Tone control, headphone stage, and push-pull output.

Long story short, the amount of 6.3 volt needed exceeded the rating of the transformer. Plus, in a RIAA stage, it is preferable to run the heaters on DC. Handily, the transformer I am using (Hammond 370FX) has a 5v winding. So I thought I could run the RIAA stage off an arrangement like this - the capacitor is 47000µF













Then, to relieve the load on the 6.3v I thought it might be possible to run 3 more filaments from this arrangement, for a total of 5 12A*7 tubes running on DC heaters.

Sadly this arrangement was not suitable – the DC voltage dropped to 5.5V which is too low to run filaments on, plus that diode was running rather warm, to make an understatement.

If I removed all load from the DC except the RIAA stage, the DC voltage was closer to 6.1 which is tolerable. But this leaves me with the following problem:












The transformer is rated at 5A on the 6.3V winding. So, my original plan was to use the DC to take about 900mA of that load. Alas this plan did not work so now I need to find around 1000mA of 6.3v ac from somewhere.

Worse, I also discovered that the tracks on the PCB I'd set up for the 6.3vac were not thick enough for the 5A load. They were dropping around 0.44v which at 5A equates to 2.2 watts of heat.

Having PCB tracks dissipating 2.2 watts of heat is a Very Bad Thing.

Clearly some thought and remedial action required!

The initial idea - swiftly dismissed - was to re-make the PCB with larger tracks. However in the end I opted to rework the connection to the EL84s so that they don't go through the PCB. That will save the board from burning up.

Only problem is that it will look ugly. It will work fine but I am not pleased.

Second problem. Where's that extra 1000mA of 6.3v going to come from?

Only one possible solution. A second, smaller, filament transformer.

As luck will have it, a transformer rated at 6.3v 1000mA was sitting in my box of spare parts. And with a small amount of re-shuffling, will fit inside the chassis.

However. It will add weight, and it will forever bear testimony to the error made in the calculation. So I am doubly not pleased.

However, the happy ending to this awful debacle is that it doesn't set the schedule back too far, and the amplifier will work completely as intended, and I do not need to re-make the board. So not a total disaster.

But definitely some lessons learned for the next project.

Here endeth the confession. 

Ending on a good note

Before the Great Filament Supply Disaster of 2018, I had the output stage and the amp board running and the B+ and other HT voltages were exactly where I planned them to be (so my high voltage design is fine, just the low voltage stuff I messed up!) and the amp was running with a signal source (=old iPod) connected, through a small pair of speakers. It sounded great, and the quick measurements through the oscilloscope with the function generator showed the response and power exactly where they were supposed to be. And there was no hum!

More later when the panels are back and more building is completed...