Solid State Bucking "transformer"

[maxhifi]

I have a small portable Japanese domestic market record player, which has a power consumption of 4 Watts. The AC line voltage in Japan is 100V, versus the nominal 120V available at the wall here in Canada. I would like to adapt the player to work here in Canada, so I was thinking about options.

I was thinking about it, and instead of retrofitting a new power transformer, or buying a heavy step down transformer, I thought hey, what about constructing a diode based voltage dropper? I sketched a circuit representative of what I'm thinking. In theory this should drop the AC voltage by the junction voltage multiplied by the number of diodes in the string, and should work well up to the maximum current rating of the diodes in the string.

For the attached example, there's 8 diodes in the string, so the voltage drop should be on the order of 0.6V * 8 = 4.8V. In order to get the 20V voltage drop I'm after, I should use about 20V / 0.6V = 33 diodes.

Now, a 1N4007 costs about $0.13 in quantity at digikey, so to build this dropper I'd need 66, so about $8.58. Not bad price wise, and it would work for any equipment consuming less than about 1A.

Any thoughts? Anyone do this before? Any reason I shouldn't do this? To think about it more, this concept should work great for vintage tube gear, and save on the weight and cost of transformers.



EDIT - One thing not mentioned above, is the fact that this is going to distort the waveform a bit. I'm not sure of how much consequence this will be. See below image for what I mean. I'm thinking that since it's only a notch and not a spike, it won't be audible.

 

[Hyperion]

Interesting, but not something I am personally considering as I would need 466 diodes.

 

[maxhifi]

I realized after posting that I did the math wrong. The diodes are going to drop peak voltage, not AC RMS voltage, so it's not going to be as clear as using 20V of diode drop to get 20V less AC. The peak voltage required for 120V is about 170V, and the peak voltage required for 100V is about 142V, so I should figure on needing enough diode drop to lose 28V, to get the RMS voltage down to about 100.

This excludes the small notch during the zero crossing when both sets of diodes are biased off. I think it's probably possible to do the math, but considering that junction voltages vary, and that I haven't calculated the area under a curve in years (to get the actual RMS voltage of the non-sinusoidal blue waveform), I think I will just figure it out experimentally, by starting with too many diodes, and removing them until the DC voltage at the record player's power supply is correct. I should dissipate about 0.8W in the diodes, so I'm not very worried about that.

@Hyperion, this won't work in the UK, the waveform distortion would be too great to get all the way down to 100V. It would however be a possible technique for using eastern bloc or grey market Chinese equipment which is designed for nominal 220V. This assuming it works - I haven't tried it, this is just the sort of thing my mind wanders to.

 

[BinaryMike]

One thing not mentioned above, is the fact that this is going to distort the waveform a bit. I'm not sure of how much consequence this will be. See below image for what I mean. I'm thinking that since it's only a notch and not a spike, it won't be audible.

It's a sharp discontinuity and it will definitely be a rich source of harmonics, but that might not matter if the downstream equipment has good rejection of powerline noise. Most audio equipment does have excellent powerline noise rejection.

 

[mondialfan]

It seems to me that it would be much easier to just change the 100V transformer out for a 120V unit. This is especially true if its only a standard single voltage secondary with CT. If there are multiple voltage taps used on the secondary then it gets a lot harder to source.

 

[elnaldo]

Did you check the original transformer? Many units can be wired for 120V inside.

 

[Watthour]

Does the unit rectify to DC as part of its supply, or simply operate on AC? If so, what is the DC supply in the unit producing? It may simply be easier to eliminate the transformer and find/build a DC supply at the proper level(s).

 

[maxhifi]

Oh, it's a very simple unit, with a step down transformer, and then a full wave rectifier with a capacitor input filter. I think it needs about 11V to operate.

At the moment I just hook it up to a variac set to 100V, and it works great. Well, it worked like it needed a new idler wheel, so I sent the old one out for exchange, and the new one should arrive soon. Once it does I'll come back to this...

Player in question is a Nippon Columbia GP-3. I cannot for the life of me find a schematic diagram for it online, and I'm usually pretty good with digging up impossible to find info. I did however find a cross reference to an Astatic stylus to replace the hard to get Japanese one, and also a substitute for the idler wheel.

I think I do want to try the diode dropper, just for fun as much as anything else. If it is too noisy, I'm not out much. The noise rejection of the internal power supply is likely okay but not great - it looks like just a simple capacitor input power supply. The transformer itself will probably filter out harmonics to some extent.

 

[elnaldo]

I'd drop the secondary voltage, not the mains. You are feeding 20% more, so output will be 20% more, perhaps you need to drop just 2 volts (3 diodes) (6 actually).

You could even drop (or add a voltage regulator) the rectified DC voltage.

 

[BinaryMike]

I'd drop the secondary voltage, not the mains. You are feeding 20% more, so output will be 20% more, perhaps you need to drop just 2 volts (3 diodes) (6 actually).

You could even drop (or add a voltage regulator) the rectified DC voltage.

That's a risky approach because of the possibility of transformer core saturation at a primary voltage that much over spec.

 

[Watthour]

If you think the load is very low (less than 30 VA) this would be the cheapest buck-boost transformer you'll ever find: https://www.homedepot.com/p/Newhous...ormer-Compatible-with-Ring-Pro-30TR/307787405

Just connect the 16V secondary out of phase with the incoming line and you should drop the line to an acceptable level - Presuming you have room for a second transformer .

 

[maxhifi]

The record player is really little, and made from transparent plastic (pic from net shows the non transparent version). It's about as small as they come, if you play an LP it partially covers the speaker. These things were kind of trendy to have in the late 90s, at the very beginning of the whole vinyl revival thing. Its claim to fame is it can play records vertically, or even upside down, with the included record clamp. It also has a pitch control. I realize it's nothing more than a toy, but hey, it's fun. They were never exported outside of Japan, hence the 100V transformer.

I thought of the diode idea as an alternative to modifying it or adding weight, since the internal workings are externally visible.

 

[maxhifi]

Here's an internal picture, also from the net (not mine but very very close)

 

[Watthour]

Well, that IS quite the little record player then. For the novelty value, it may be best to simply stash it away and use a buck-boost when you feel like playing with it, instead of altering it at all. .

 

[maxhifi]

Well, that IS quite the little record player then. For the novelty value, it may be best to simply stash it away and use a buck-boost when you feel like playing with it, instead of altering it at all. .

I've been using my EICO variac, which works great, but requires removing it from my workbench. I was planning on putting the diodes on to a piece of perf board, and then mounting that into a plastic enclosure as part of a custom extension cord, or something like that. I mostly just want to see if it works. As binarymike alluded to, the diode fix will keep the voltage below transformer saturation, but MAY create noise on the audio side.

This record player has holes in the back to allow for wall mounting, and can play records when set up that way. I've had it on the wall as decoration for years.. decided to make it work properly recently.

 

[Binkman]

Well, that IS quite the little record player then. For the novelty value, it may be best to simply stash it away and use a buck-boost when you feel like playing with it, instead of altering it at all. .

It is of only 'curiosity value' I'd leave it be and get on to the stack to post.
I'm really bored

 

[BinaryMike]

If the diode stacks don't work out, you could just bypass the internal transformer-rectifier stuff and try a conventional unregulated 9VDC wall wart. They deliver about 11VDC when loaded well below their current ratings. A modern 12VDC regulated wall wart might be suitable as well. If 12V is too high, it could be dropped into the desired range with a few forward-biased diodes in series. Does the motor have an operating voltage marked on the sticker?

 

[petehall347]

to keep it original looking a substitute transformer should do it and shouldn't cost much to do .

 

[maxhifi]

It is of only 'curiosity value' I'd leave it be and get on to the stack to post.
I'm really bored

I clicked buy it now on 100 1N4007 diodes for $1.13 including shipping on the slow boat from China. Will update once they arrive, and take some photos of the waveform, and measure the THD, then check to see if it really makes any noise or not.

Voltage on motor is marked 9VDC, it's got an internal speed regulator and can take something like 6-12V. Of course at 12V it's harder on the regulator. The thing can either take six batteries (9V), or running on line power it sits at about 11V for a little more jam. Once the new rebuilt idler wheel comes back, I will get an idea how much potential it has, I might try and soup it up a bit.