Not really surprising if you think about it. The transformer always drops the same number of volts at its rated output current, but regulation is specified as a percentage of the nominal voltage.
There is no physical effect whatsoever, it's all a matter of accounting.
Let's take an example: A 240V:24V transformer rated at 240VA (ie, 24V, 10A output) and 5% regulation. The internal impedance will be 5%. In absolute terms referred to the secondary, that is (24/10)*(5/100) = 0.12 ohms. The volt drop is 1.2V and the no load output voltage 25.2V.
Now let's use the same transformer on 120V. The no-load voltage will be 12.6V and the voltage at 10 amps will be 11.4V, since as we stated above, the transformer's internal impedance drops 1.2V at 10 amps. The regulation is now (12.6-11.4)/12 = 10%.
If we expected it to still deliver 240VA at 120V, we would be asking 12V at 20A out of it, and the regulation would be (12.6-10.2)/12 = 20%.
Let's take an example: A 240V:24V transformer rated at 240VA (ie, 24V, 10A output) and 5% regulation. The internal impedance will be 5%. In absolute terms referred to the secondary, that is (24/10)*(5/100) = 0.12 ohms. The volt drop is 1.2V and the no load output voltage 25.2V.
Now let's use the same transformer on 120V. The no-load voltage will be 12.6V and the voltage at 10 amps will be 11.4V, since as we stated above, the transformer's internal impedance drops 1.2V at 10 amps. The regulation is now (12.6-11.4)/12 = 10%.
If we expected it to still deliver 240VA at 120V, we would be asking 12V at 20A out of it, and the regulation would be (12.6-10.2)/12 = 20%.
Last edited:
Hi Dan. 55 Hz causes a beat frequency that often can be heard if regenerated. My turntable used 52 Hz and 70 Hz. Being only 25 watts the problem is not too bad . The output transformer being a useful filter. Why 52 Hz? 50 Hz was used with a 5 mm shaft and 6.33 mm pulley. Going to a 6 mm shaft meant no pulley. Circa 52 Hz. No pulley = no extra wow and flutter from machining errors. The shaft size must not go above that too much as it causes drag.That is a small but unwelcomed problem, it causes the stable speed point to be delayed when a hysterisis motor ( heat, drag, time, approximately 1% speed change cold to hot). Some drag is useful, it acts as damping.
I was given the Garrard heritage to preserve. In doing it I became a convert. I like like hysterisis motors. I think the 1% warm up is small price to pay for the elastic magnetic field effect. That is the rotor gradually slips further behind the stator field as load is applied. Held at a sensible level it makes for very low vibration. This also makes the motor show the positive effects of a low distortion PSU. 1/3rd, 1/5,1/7, 1/9 th harmonics cause the elastic field to modulate. As a turntable is much like a loudspeaker the stylus picks this up ( motor and shaft the coil and the platter the cone ). Reduce this problem to 2nd only and great performance can be had.
Taking the whole system I find - 52 dB to be a practical distortion level for the Garrard/Thorens hysterisis motors. Most of this is generated inside the motor magnetic circuit. Inside a Linn motor this is more like - 26 dB. I suspect this must be true for DD as the way they work is similar. Some say pre-distortion can change that. I have doubts about that. The motor notches as it is rotated by hand. Hard to think one can produce anti- notch pre-distortion. Not least the way it works is the way it works.
The PSU I made as a variac just about addresses some of these distortion problems. When I order parts I will make a complimentary feedback pair version with TTC 5200 TTA1943 and MJE340/350. It should be nearly as good as full regeneration. I plan to slug the transistor 1 cb as I have no need for bandwith. Was dam glad I had the Variac today. I was doing PSU regulator drop out. 180 V perfect, 177 V dreadful. 178.6 V OK ish ( - 100 db hum, - 120 dB when low ) . The regulator is 13.4 V in and 12.25V out. It is a 3 pin TO220 device of 1.25 A with no readable markings. Low noise and drop out in 7812 format. Try as I might I can not find anyone who claims to make one. The noise beats the 0.003% of the 78 series. I must say is takes much effort to beat a LM317 with a capacitor filter to the reference pin. I used 470 uF and got a bit extra. Took a while for the voltage to come up. No obvious problem. I used the usual 240R upper gain setting arm.
Taking the whole system I find - 52 dB to be a practical distortion level for the Garrard/Thorens hysterisis motors. Most of this is generated inside the motor magnetic circuit. Inside a Linn motor this is more like - 26 dB. I suspect this must be true for DD as the way they work is similar. Some say pre-distortion can change that. I have doubts about that. The motor notches as it is rotated by hand. Hard to think one can produce anti- notch pre-distortion. Not least the way it works is the way it works.
The PSU I made as a variac just about addresses some of these distortion problems. When I order parts I will make a complimentary feedback pair version with TTC 5200 TTA1943 and MJE340/350. It should be nearly as good as full regeneration. I plan to slug the transistor 1 cb as I have no need for bandwith. Was dam glad I had the Variac today. I was doing PSU regulator drop out. 180 V perfect, 177 V dreadful. 178.6 V OK ish ( - 100 db hum, - 120 dB when low ) . The regulator is 13.4 V in and 12.25V out. It is a 3 pin TO220 device of 1.25 A with no readable markings. Low noise and drop out in 7812 format. Try as I might I can not find anyone who claims to make one. The noise beats the 0.003% of the 78 series. I must say is takes much effort to beat a LM317 with a capacitor filter to the reference pin. I used 470 uF and got a bit extra. Took a while for the voltage to come up. No obvious problem. I used the usual 240R upper gain setting arm.
This is just to ruffle Nige's hair a bit - a (gulp!) direct drive Thorens! 
To be honest, this model did not live very long, 3-4 years at best. It seems Thorens was by then so deeply associated with belt drive that nobody wanted a direct drive from them, leaving such audio crimes to the likes of Technics, Sony, Denon, et al.
I mean look, who but the lowly wants a turntable with no fuss, right?
To be honest, this model did not live very long, 3-4 years at best. It seems Thorens was by then so deeply associated with belt drive that nobody wanted a direct drive from them, leaving such audio crimes to the likes of Technics, Sony, Denon, et al.
I mean look, who but the lowly wants a turntable with no fuss, right?
Attachments
Look carefully and you will see the rumble is only so-so. Very good when one knows the problems the motor brings. Picture this. The motor spins far too show to be working at it's best. Then there is the speed regulation which is as good as it's imformation the servo gets. The motor be it fed with sine or square waves makes a crude sine wave out of the current given. Sorry Denon the input wave is not a big deal. The distortion is about -18 dB down and looks like a sine wave pushed through a gap. It looks like a triangle wave with a long nose on it. The harmonics are vibration. Then there is cogging which the flywheel effect helps. The bigger help is the way the servo works. If it is a lazy servo it allows more wow and less flutter. I always notice wow and flutter is very low suggesting that wasn't the route taken. Having said that I have a very domesticated JVC DD. It sounds crude but can put the LP12 to shame at times. The Thorens TD145 is more put to shame as it slightly comes 1/3 of the group. The JVC is like someone you would date for one reason alone. Sad to say that reason is an important one. The JVC has verve. BTW JVC hold many patents on DD. Outside of Japan they never show how good they are.
I put up a question about a mystery 3 pin TO220 voltage regulator that matches my reference device I made. It has one advantage over that in having built in current protection. Whilst building that in is no big deal the noise it would generate might be. If not careful a bypassed LM317 would be better.
I think I have worked out what the mystery device is. LM340T12.
Now my question. The output noise of the 12 V version is 75 uV and the 5 V 40 uV. By gain adjusting the 5 V version as if a LM317 Vo = Vr x {1+( R2/R1)}. One presumes nothing gained by doing that as 12/5 = 96 uV . The gain might be the noise of Vref is presented to the usually grounded pin. Maybe just maybe the noise can be shunted to ground and more so than the usual 40 uV output noise? The devices are cheap so I can play with them. I will post if interesting. I will use LM317 also as it is very good with an oversized noise decoupler ( 470 uF ).
One thing I discovered. Slightly lower grade caps work best at the output. Some would use a ceramic and cheap electrolytic together. The old tantalum would not be ideal as often thought. More so as they fail. Mr Marsh was talking about transient performance. As far as I can tell oscillation is the same problem. LT advise a 100 kHz square wave device as a test. A TDA 2030 could do that into 8R perhaps?
I think I have worked out what the mystery device is. LM340T12.
Now my question. The output noise of the 12 V version is 75 uV and the 5 V 40 uV. By gain adjusting the 5 V version as if a LM317 Vo = Vr x {1+( R2/R1)}. One presumes nothing gained by doing that as 12/5 = 96 uV . The gain might be the noise of Vref is presented to the usually grounded pin. Maybe just maybe the noise can be shunted to ground and more so than the usual 40 uV output noise? The devices are cheap so I can play with them. I will post if interesting. I will use LM317 also as it is very good with an oversized noise decoupler ( 470 uF ).
One thing I discovered. Slightly lower grade caps work best at the output. Some would use a ceramic and cheap electrolytic together. The old tantalum would not be ideal as often thought. More so as they fail. Mr Marsh was talking about transient performance. As far as I can tell oscillation is the same problem. LT advise a 100 kHz square wave device as a test. A TDA 2030 could do that into 8R perhaps?
I think I will correct my idea . The capacitor of a LM317 is not shunting the noise at reference pin. The capacitor causes the output impedance of the regulator to be lowered. Thus the noise and impedace curves look very similar when LM317 and other devices. LM317 being a 1.25 V device might have a better chance of low noise.
One idea is to raise the resistance of the feedback arms. 1K and 10K for example. It might be true that the capacitor can be smaller which avoids the mid range frequency hump the cap gives at about 2 kHz. If the lower arm repalced by a zener and retaining the cap it might be best possible?
One idea is to raise the resistance of the feedback arms. 1K and 10K for example. It might be true that the capacitor can be smaller which avoids the mid range frequency hump the cap gives at about 2 kHz. If the lower arm repalced by a zener and retaining the cap it might be best possible?
Begging your pardon Nige, but please show me a better spec for a TT anywhere in its price range? By anybody, any time?
Start with Thorens products, avoidig only those models which cost 2 or 3 times as much.
I'm just being picky over Thorens DD.
Funny you should mention Magnox reactors. Collleen took me to Sizewell A and B for a picknic. A was a Magnox and B a water cooled job. Hinkley might get a new one. The French will be happy as the UK will keep the skills alive to build them. I told the PM we need 54 extra ones to power electric cars. Perhaps the same on top. He is my MP so has to put up with me. I hate nuclear, I hate the others more.
Was the Dual better? Is it the same motor ? Dual would do that if asked. My friend Martina was parts lady at Thorens, she was family friend of the Dual people. The Black Forest people are that close.
What am I building ? A voltage regulator better than a carefully set up LM317. Tougher than I thought. I suspect a zener diode + cap ( 6.2 V ? ) , op amp, V adjust , power transistor, oven , and crude current limiter is the only way to beat the noise levels. The crude limiter is two of the H4 halogen bulb filiments. The H4 can also be part of two x 10 000 uF in Pi filter arrangement. I suspect NE5534 is a good choice as it will power the TTC5200 ( 2SC ). 2 amps /50 = 40mA. A pair in a 5532 should do it and slightly cancel the noise ( 2.5 nV). To be frank zeners that were said to be so bad when young are better than expensive bandgap devices. LT's better chips are zener.
LM317 is 0.003% noise ( as are most 78 and posh versions ). If the gain device is filtered ( bootstrapped) it is 0.003% of 1.25 V. that is 0.00032% of 12 V. 37.5 uV ( 6 dB better than the better fixed voltage types of typical 75 uV ). From what I worked out TL431 and complimentary feedback pair could beat that. TL431 and op amp + TTC5200 looks possible.
Funny you should mention Magnox reactors. Collleen took me to Sizewell A and B for a picknic. A was a Magnox and B a water cooled job. Hinkley might get a new one. The French will be happy as the UK will keep the skills alive to build them. I told the PM we need 54 extra ones to power electric cars. Perhaps the same on top. He is my MP so has to put up with me. I hate nuclear, I hate the others more.
Was the Dual better? Is it the same motor ? Dual would do that if asked. My friend Martina was parts lady at Thorens, she was family friend of the Dual people. The Black Forest people are that close.
What am I building ? A voltage regulator better than a carefully set up LM317. Tougher than I thought. I suspect a zener diode + cap ( 6.2 V ? ) , op amp, V adjust , power transistor, oven , and crude current limiter is the only way to beat the noise levels. The crude limiter is two of the H4 halogen bulb filiments. The H4 can also be part of two x 10 000 uF in Pi filter arrangement. I suspect NE5534 is a good choice as it will power the TTC5200 ( 2SC ). 2 amps /50 = 40mA. A pair in a 5532 should do it and slightly cancel the noise ( 2.5 nV). To be frank zeners that were said to be so bad when young are better than expensive bandgap devices. LT's better chips are zener.
LM317 is 0.003% noise ( as are most 78 and posh versions ). If the gain device is filtered ( bootstrapped) it is 0.003% of 1.25 V. that is 0.00032% of 12 V. 37.5 uV ( 6 dB better than the better fixed voltage types of typical 75 uV ). From what I worked out TL431 and complimentary feedback pair could beat that. TL431 and op amp + TTC5200 looks possible.
- Status
- Not open for further replies.