Sound Quality Vs. Measurements

Status
Not open for further replies.
On the other hand, things at ON Semi are smokin'!

Take a look at MJL4281 and its complement, 4302. They have increased the voltage by 100V now at 350V, dissipation, up from 200 to 230W, and ironed out the current capability no end. Its Ft is also up from 30 to 35 MHz.

Comparing the data sheets with MJL 3281/1302, which I have like 26+26, its impulse power is up from around 5 to around 7.5 amps.

I've got to get me some!
 
I thought I went to extream lengths to find out things. To build a NE5534?
Just for Fun - Discrete NE5534

Talking of extream lengths. As soon as some heat washers turn up I am going to attempt my most likely to be doomed amplifier yet. This is based on a Blomley amp. A TL071( 1/4 TL074 ) will feed into a complimentary double VAS both in common base. These will be biased to 0V via 2 x 1N4007. The bias current via 100K and 15V+/-. The most likely transistors MJE340/350 for now. 100K is almost no bias, goodness knows what will be used. The rail will be +/- 53 V and VAS resistors 5K1. Output stage is complimentary pairs acting as super gain singles ( 0.55V start of turn on ) . It is more usual to use a CCS to the divided VAS's. As only 50 Hz will be interesting to see what it does.

The big nasty problem is integrating an op amp into a power amp feedback loop. We are not talking slewing. This nearly always is instant fire. Now to compound things I will have to use an inverter stage as common base VAS's will the invert phase of the output ( 1/4 TLO74 as inverter ). I could and should use shunt input feedback which would help. If I ever get to fitting the output transformer the phase is solved as it is then whatever phase I choose. I will set that up as 115-0-115 as that means conversion to 0-115 is easier. I found in the past it helped as it cancelled the harmonics I wanted to cancel. Slight down grade when 0-115 if so. I have made some very fancy linear inverters. Often allowing me to research the amplifier for my own interests. This time I want to take it back a step. This amp if it works will doubtless have many blow ups. If this could be limited to TL074 and outputs it might make for quick repairs. So far this year I have ruined 3 Variacs. Mostly without being too stupid. To ruin a repairable amp would be OK.

The oscillator is built following an old RA Penfold design . 1/4 TL074 RA53 thermistor output to -ve input( 5K ) and 680R -ve to 0V. Feedback 10n1 and 316K in series output to +ve input and 10n1/316K +ve input in paralell to 0V. Took me as long to type this as to build. 0.1 % distortion 1Vrms out. If you find a RA53 you must take it away from heat. If so a very practical way to do it. In the past with little effort I found the distortion can be reduced. 10n1 was hand picked from very cheap 100 ppm polyester. The 316K from my RIAA amps ( MRS25 value ). The measured frequency 49.68Hz. The meter 0.002% accurate. This shows the value of JFET inputs. The calcualted and actual are very close and allow cheap components to work. My guess is the JFET is near to zero error. Strangely the fantastic frequency meter is also the rubbish signal generator! I did have a Racal one with Nixie tubes which was even better. It died and no circuit diagram.

Dejam. RA53 seems to be USA origin made in out cider country! Bowthorp. There was also RA54, never tried one.

The purpose of this device is so as to be able to accurately measure things without a true RMS meter. Even if I had one it wouldn't help for the design I am working on. I will build an opto Schmidt from the mains. Via a 6 pole filter that will give a very stable 50 Hz sinewave. UK mains is very frequency stable. The voltage stability would out rank the RA53. Opto gets rid of all the safety problems.
 
Last edited:
nigel pearson said:
UK mains is very frequency stable.
Only when averaged over a day. The 'instantaneous' frequency can vary by 0.1-0.2Hz as generation tries to match the load.

See U.K. National Grid status for current values. The frequency just moved from 49.955 to 50.055Hz over about a minute, so either a big load just dropped off or a new generator came on.
 
My friends in Germany say +/- 1 Hz. I doubt it.

Using polystrene caps I managed to keep a Sony clock-radio correct to 6 seconds a day using a RA Penfold bridge. With NPO ceramic perhaps a tad better? My 49.68Hz is excellent using what was to hand with calculation alone. The application is 50 or 60 Hz so I can live with it. I will correct it as it is 5 minutes work. I will correct the series RC ( 10.1nF and 316K) as it is easy to do it. The Penfold cirfcuit works well with one side slightly different.

If my phone is right as a calculator ( and me ) about 263K gives 60 Hz. 1M580 or 1M5 in paralell would be close.

For 45 RPM on 33 1/3 50 turntable about 234K or 903K ( 910 K ) in paralell.
 
Exactly. I suspect 51Hz was taken from an old analogue meter that was moving away from 50 Hz ( 50.3 Hz ? ). Even so unlikely.

People say the flat top distortion seen on the mains is due how power is drawn. This shows a 100 kVA diesel genertor put in at the end of my road to supply us when a circuit was being ungraded. There is next to nothing in our street as a load. It seems to me the flat top might be a simple form of regulation seen both on this and the National Grid.? Note how the grid is slightly better. Was very luckey to get this, rushed back from the pub as the guys said until midnight. At 10.30 PM we went back to National Grid. The voltage is potted down to protect the scope. The scope is opto coupled so is direct and not transformer isolated.

uyJ5Fpo.jpg
 
Dejan. That is a lovely amplifier you sent me. Thanks. You will see a very cowardly amp today I suspect. I realised I need this thing to work and not play hero's. I will have 60 Hz and 67.5 Hz. This will protect the Thorens TD145 belt when playing 45's. It's my last new genuine Thorens belt and 45's kill them in one month. If the LP12 was ever better it is it's belt. It is hard to say why as it is no big deal different. I suspect Thorens found by making it smaller it had slightly better wow and flutter. When Logic turntables measured belts the Thorens was best of what was easy to buy. B&O the very best. I see B&O bad mouthed a lot. I would have loved to have worked for them. They certainly were minimal even by my standards. I think they crossed the line on that. No other company pushed out so many ideas. No other company used so little of it's own research . Dolby HX was B&O. The B&O MMC 20 was the most interesting pick up. The B&O researchers studied MC pick ups. At first they assumed people liked distortion. Shure had promoted this idea so as not to pay royalties to Ortofon to simplify the story. B&O found that all MC had far better stereo separation and linearity. They solved the problems in MMC 20. Grado and Ortofon had arrived at the same conclusion.

Capacitors and batteries. I have a 4.5 V niCad in the central heating timer. After 20 years it has died. I was thinking to use a 2F2 5 V cap. My concern is what series resitor use to prevent damage? I was thinking 100 K.

I notice hpeter you use DT990 Pro. My Beyer rep told me DT990 was the prefered version as DT990 pro was for discovering recording faults. DT990 was slightly less good drive units with a filter. Mr House ( real name ) was not a technical man so who knows? To 25 kHz filter them is no big deal. I seem to remeber that was what it was?

Screen drive , is that to g1 or g2? EL84 would be my choice if so. I am trying to get JJ to make a TR34, that is a special EL 34 with no g2g3. I suggested to her that g1 might need to be in a different place. The idea is a cheaper PX25 with indirect heater. It would work for most people in the same socket as EL34. If she listened I have a Marshall amp to try. A Fender like sound in triode? Shame not like GU50. They can be pulled out instantly.
 
Member
Joined 2002
Paid Member

Attachments

  • EN 1.JPG
    EN 1.JPG
    74.3 KB · Views: 167
  • EN 2.JPG
    EN 2.JPG
    58.6 KB · Views: 166
Last edited:
Very interesting . When at my old company we asked for voltages seen arround the world. Not the internet, what our technical people knew first hand. Approximately 180 to 280 V in the 220/230 areas. I could well beleive the frequency would also swing vastly when dropping to 180. The point the people made was local arrangements sort it out and circuit breakers keep things safe if over current. Better to have some elctricity than none. The most suprising was Austrailia sometimes hitting 270V. I wonder with that one if a myth ? My friend Martin said so 30 years ago and an engineer/hi fi dealer said so recently. At my now company I use 180 to 280 when I can. A friend in the USA says 130V is not uncommon.
 
I measured 275V in a monastery in rural Armenia after a fuse blew on my mic preamp in a recording session.
Fortunately, I had a spare and had an assistant who had a massive (about 40kg) Russian power transformer with a multi-tap secondary and employed that after the fact to bring the voltage down. Apparently that sort of voltage was not uncommon and that's why he brought the transformer along. He left it in the bus because he didn't want to carry it out unless it was necessary! Good thing I had a spare fuse.
 
Member
Joined 2002
Paid Member
Apparently that sort of voltage was not uncommon and that's why he brought the transformer along

I had tested a nice manual adjustable USSR transformer for domestic use, output +15%, -10% rel input.
1kW (continuous) double C construction, 220V or 120V input selectable, output voltmeter, strong black bakelite housing.
The slide adjuster was sliding over both primary and secondary windings (constant impedance?)

George
 
......................
When 230V was adopted across Europe, stated tolerance was +5%, -10% (207V-242V) for backward compatibility
.................
No.
The European Harmonised voltage range is 216Vac to 253Vac
i.e. 230Vac +10% ~-6%

Prior to harmonisation, the UK used 240Vac +-6% i.e. 225.6Vac to 254.4Vac
I think that many assumed the old UK standard was 226 to 254 and I have seen many documents stating these values.
 
Last edited:
There is also a volt drop allowed in UK house wiring. These days it is not a realistic figure as the full loading of 32 amps is unlikely. In the past electric fires were more common, My kettle knocks it back a bit.

In the UK we use Ring Mains. It is along time ago I studdied this so forgive any errors. By having fuses in our plugs we share the circuit between many outlets. The single use current for 2.5 mm cable is about 27 amps in older assumptions of buildings codes. As it is a ring we can easily get to 32 amps ( 54 ). Wise as seldom do 32 amp fuses blow at even 40 amps short duration, Each socket is allowed a spur of one set of double sockets. Unlike the USA this is not to power lights. It allows a double socket of 26 amps. This is realistic as the cable will not get hot even above full load. The better thing is we can wire a house at lower cost and have lower resitance on the places far away from the fuse board. Hi fi people often have dedicated supplies for the hi fi. Somehow the ring main is a no no in the Folk law. This is nonsense as a dedicated ring main is ideal . Our lighting is a Radial system ( don't know the proper name ). It is 6 amp fused on cable of > 11 amps ( 1 mm ). New regulations concerning insulation has come in with a 25% derating if I was listening to an electrican friend correctly. As luck has it that still works in the newer regulations. The insulation guys do not know this and care even less. It was far sighted to use 1 mm cable. The ring main has one big problem. A single socket pair can be disconnected from one cable and still the Ring works. A little unlikely as for one to be loose that should mean two cables loose. The 27 amps is still OK as it is a conservative rating. We are allowed more than one set of sockets on a spur if fused. This is like an extention cable wired in. We have a neat single socket size plate with fuse holder ( 1 inch type ). Single sockets are OK, they sometimes are used to take up less space.

I am told the EEC would like us to change the the European system of single socket single cable to the fuse board. I hope not as the UK system is so logical.

When the 230 V standard came in RS the component people sent me a very good report. It showed how the standard would mean no change and was just an easy one fits all way of saying things. It was remarkable how in picture form the standards were very similar. Often equipement was 220/240V. The makers had made sure it could work universally. I guess that says it best?

At work the kettle was rated at 10 amps. I fitted a 6.3 AT fues to test a theory. The staff knew this. After months of testing we gave up. The 4 minutes or whatever never exceeded the fuse capacity and didn't seem to degrade the fuse. When fitted to amplifiers they often go when 5AT. It must be a very big overload.

One daft book said " Were it not for Ring Mains the Worlds supply of copper would be used up by now". Silly man, we are the only ones who use it.

I was forced to do electrical engineering at college if wanting to do electronics. Glad they did that as although not my day to day interest isome of it went in.
 
How can frequency vary? The grid is just that, a network of sources and loads. All of the sources have to be line synchronized or things go very wrong.


132VAC is not uncommon in the US. I've seen it numerous times.

That will go in the book. Anyone know typical lows ? Outside USA also. Japan is said to go to 90 V ( 100 V system of 50/60 Hz ).

The frequency is odd. Over a long period perhaps. The mas of the turbine is many 10's of tones ( 100 's ? ).
 
TheGimp said:
How can frequency vary? The grid is just that, a network of sources and loads. All of the sources have to be line synchronized or things go very wrong.
Yes, an AC grid is all at the same frequency. The frequency varies as the load has a slight temporary mismatch to generation. Too much load and the generators slow down. Too little load and they speed up. The control systems notice this and add more/less coal/gas etc. to produce more/less steam and so the generator can speed up/slow down to bring generation exactly equal to the load again. This process is helped by the fact that higher speed can mean higher voltage so higher load power, and some loads vary with frequency too.

The reason it all works so well is that there is a lot of stored kinetic energy in the rotating machinery and this helps smooth things out and gives sufficient time for the control system to react.
 
Status
Not open for further replies.