You are reading the older HTML site

Positive Feedback ISSUE 17
january/february 2005


Shuguang 845M tubes
by Roger S. Gordon


In PFO issue 15, I reviewed the 845B vacuum tube that had just been released by Shuguang in China. At the end of the article I mentioned that Shuguang was working on a metal-plate 845 tube, and that they expected to release it in September 2004. As it turned out, there were design problems with the 845M (as the metal-plate 845 was designated), and it took a number of months to get the problems worked out. The first batch of 845Ms was released in late January 2005, and I obtained a matched pair for my deHavilland 845-G SET monoblocks.

Since metal-plate 845s have not been made since the 1930s, it is not surprising that Shuguang ran into problems. A lot of technology can get lost in sixty years. The reason metal-plate 845s were replaced by graphite-plate 845s is not known, but it was probably due to higher production costs, as the metal-plate tubes require more welding and fabrication. In addition, metal-plate 845s produce 20 to 25 percent less power than graphite-plate 845s. Sixty years ago, there were not too many audiophiles. Who else would pay more money for a less powerful vacuum tube because it sounded better? If you are interested in the technical specifications of the 845M please refer to Ed Sawyer’s excellent article at

The search for metal-plate 845s by audiophiles has been likened to the search for the Holy Grail. Not very many were produced, and very few survive. Most of the survivors no longer function due to broken filaments, gas leaks, and other problems, but the few working ones have created a legend. Metal-plate 845s are said to sound better than all other 845s, but because they are unobtainable, there has been no way to determine if the legend is true. With the appearance of the Shuguang 845M, we can put the legend to the test.

I invited two of my audiophile friends over to assist me in my comparisons. One of them was kind enough to bring over two pairs of NOS RCA 845s and a pair of United 845s.  The RCAs and Uniteds are considered the best sounding of the obtainable 845s, though of course, obtainable does not mean cheap. A pair of good-quality used United 845s currently sells for $1400, and NOS RCA 845s sell for over $1000 a pair. Our procedure was to play the same recording again and again—Louis Armstrong’s “St. James Infirmary,” a Classic Records reissue on 45 rpm vinyl. First we listened to the 845Ms, then the 845Bs, then the 845Ms, then the Uniteds, then the 845Ms, then the RCAs, then the 845Ms.

The tube swapping took some time, as the bias voltage had to be reset when each pair of tubes was inserted. We used a bias voltage of 725mv for all of the tubes. The RCA, United, and 845Bs all used approximately the same setting, but the 845Ms needed a much higher setting to reach 725mv. In addition, the bias voltage of the 845M drops by over 100mv as the tube warms up. For the 845Ms, I set the bias voltage at 825mv at startup. As the tubes warm up, the voltage drops down to 725mv. I have inspected the metal plates at bias voltages up to 850mv to see if any red spots appear, which would indicate that the plates are getting too hot. I have not seen any.

How did the tubes sound? The RCAs had the biggest soundstage, but were the least dynamic. The Uniteds were the best all-around performers. The 845Ms were the most dynamic, but we felt that they emphasized the treble to a slight degree. Interestingly enough, the metal-plate GM-70 (big brother to the 845) is also reported to have a slightly rising top. A rising top is not a major problem because it is euphonic, and can give music a more spacious, airy sound. The 845Bs had a more even frequency response than the 845Ms, but we found the 845Bs more hi-fi-ish and the 845Ms more natural, even with the rising top. We all preferred the 845Ms over the 845Bs. The difference between the 845Ms and the 845Bs was about the same or slightly less than the difference between the 845Bs and the stock Shuguang 845s. Two of us preferred the 845Ms to the RCAs because of the 845Ms’ superior dynamics, but the third listener preferred the bigger soundstage and silky smoothness of the RCAs. After the listening comparisons, we played some heavy-metal LPs at high volumes, and the 845Ms seemed to be running out of steam. There was no clipping, but they just could not play as loud as the 845Bs. Shuguang states that the 845M has only 75 percent of the power output of the 845 and 845B. Based on this experience, that appears to be true.

When purchased through, the 845Ms sell for $280 per matched pair plus shipping. The price includes having Super TNT personnel hand-select the tubes, burn them in for 24 hours, measure and match them, put them into an amp and listen to them for half an hour, then pack the tubes very securely in a foam-lined, gold lettered, hinged, hard cardboard gift box. This is quite a bargain, and can only be done at this price because labor in China is so inexpensive. 

I am going to use the 845Ms as my main tubes. I will keep my 845Bs for those special occasions when I really need those six extra watts of power—like when I play Rammstein or the six-LP boxed set of Metallica that I recently purchased. For jazz, orchestral, or popular music, the 845Ms should do just fine. They are highly recommended. Roger S. Gordon