Tube rise time and slew rate

[gain-wire]

Good day(or evening) all,
I was recently thinking about the importance of rise time and slew rate when talking about Vacuum tube characteristics. I noticed that no datasheet I ever looked at ever had this written, whether it be triode tetrode or pentode. I figure this was not an important spec back in the time, I'm guessing mainly because tubes were not used to generate square waves and pass high-speed data. that is why transistors replaced tubes in the data and digital domain. So OK, I know I am getting a little carried away from the point of my question, which is:
Do Tubes compress "softly" (rounded edges on each side of the peak) because they have a slower rise time (slew rate) than transistors? And this, combined with the fact that they operate at much higher voltages? (longer time to reach the peak since its slope is more gradual, as opposed to transistors, which have a more abrupt slope).
I'd like your input. Need input! (short circuit)
Thank you.

[richwalters]

Suggest have a look at vacuum tube science, esp Nuvistor applications . There's no limit to frequency and innovations. There's alot of RF tubes about, still used in broadcasting. Sine, square wave or radar pulse not an issue. In fact preferred.
Pentalabs is another source and loads of tubes still used in amateur rigs.

richj

[Miles Prower]

Originally posted by gain-wire
Good day(or evening) all,
I was recently thinking about the importance of rise time and slew rate when talking about Vacuum tube characteristics. I noticed that no datasheet I ever looked at ever had this written, whether it be triode tetrode or pentode. I figure this was not an important spec back in the time, I'm guessing mainly because tubes were not used to generate square waves and pass high-speed data.

That spec was never given since the VT, unlike the BJT, operates equally well at DC as it does at higher frequencies. Just about any VT will operate well into VHF. Back in the early 1970s, there were CB linears that used Class AB 50C5s to convert those 100mW "base stations" like the Archer into full 5.0W CB rigs. You won't find any specs for 11 meter operation of 50C5s since they were never intended for RF useage, but they work just fine.

For VTs, the limiting factor for slew rate is input capacitance and how fast you can charge it so that the grid sees the true signal level in a timely fashion. Slew rate becomes a problem if you try to drive the grid(s) of the final with something like a 12AX7 drawing plate currents on the order of a milliamp. It isn't going to be able to source the required current. Yet you see that all too often. I include grid drivers in my designs for just this reason.

that is why transistors replaced tubes in the data and digital domain.

No. The main considerations are ease of miniaturization and efficiency. These latest Intel chips brag about containing 450 million transistors. Just try duplicating that with VTs: get 225 million 12AU7s (were that many ever made in the whole world?) It'll take 425.25MW just to light 'em up. Not exactly a practical proposition, is it?

Do Tubes compress "softly" (rounded edges on each side of the peak) because they have a slower rise time (slew rate) than transistors?

No. Slow rise times are a problem with transistors because the device's higher inherent gain leads to some hellacious Cmiller and the construction leads to higher Ci's. Larger capacitances require bigger currents to charge them rapidly. Transistors clip hard and fast since the saturation voltages are so low that they can operate until they hit the DC rail. VTs, even at full saturation, still have considerable voltage across them, and approach saturation more gradually as cathode space charge is depleated. Being a solid, there is no lack of conduction electrons available for transistor conduction.

And this, combined with the fact that they operate at much higher voltages? (longer time to reach the peak since its slope is more gradual, as opposed to transistors, which have a more abrupt slope).
I'd like your input. Need input! (short circuit)
Thank you.

Makes no difference. Rise time is the time it takes to go from 10% of Vp to 90% of Vp.

[richwalters]

Echoing Miles comments, about a decade back a well-known manufacturer introduced the CoolMos, a high current high voltage MOSfet. The specs looked glory good, Low Rds on etc, to ramp against the competition but for switching power supplies which it was intended, there was a drive penalty. I remember r &d'ing it for a replacement. The internal gate impedance restricted the upward freq. So solid state doesn't and cannot give all the goods.

The seed of the KT88 was sown in the 12E1 with a top cap. Again a transmitting tube as the 6L6 family was an 807 also with a topcap. This was all proven stuff ! but I stand to be corrected.

richj

[Eli Duttman]

Speaking of KT88s, until New Sensor "reissued" the Gold Lion, the most affordable way to get the MOV sound was to use TT21s. The TT21's guts are the same as the KT88's, but the plate is brought out to a top cap, instead of the Octal base. The top cap plate connection allows for higher B+ voltages.

Early on, the HF performance of BJTs left much to be desired. As a youngster in the late 1950s, I had to fight to get a 3N25, which had a 250 MHz. alpha cutoff freq. It had finally become possible to build an all SS FM receiver. Cost me $12, which was serious money back then. FWIW, the 3N25 is a (sic) tetrode PNP BJT.

Slew limiting is an issue in audio, especially when loop NFB is employed. Pick a high gm type to resist slew limiting.

[gain-wire]

Thank you all! It took me good while to understand all you said... I don't know a whole lot about tubes, except I really love them! I own a handful of vintage Mullard, Dynaco and Telefunken branded 12AU7s and am hoping to design them into some audio project (or guitar amplifier) and really take profit of them.
Eli, I am still scratching my head over the Slew limiting in audio, especially with Negative feedback. I don't even know where to start taking a guess... Why is it important?

[Eli Duttman]

Eli, I am still scratching my head over the Slew limiting in audio, especially with Negative feedback. I don't even know where to start taking a guess... Why is it important?

You have a HF error correction signal, when loop NFB is employed. The circuitry must react quickly to the error correction signal. Otherwise, the "smearing" that many people object to will occur. Quick reaction is the abilty to slew. Slewing requires high gm.

[18thell]

Tube is function on EG - IP curve, EG is grid voltage and IP is plate current. Transistor is function on IB - IC curve, IB is base current and IC is collector current. Fet is function on IG - ID, IG is gate current and ID is drain current. The three element is function at a different concept. One is voltage element and the rest are current element.

Slew rate usually consider in IC speed, Transistor consider rise time and fall time which is Ft Tube consider reluctance unit in Siemons

[Eli Duttman]

Fet is function on IG - ID, IG is gate current and ID is drain current.

INCORRECT! Like tubes, FETs are high impedance, voltage controlled, devices. The transconductance of FETs is listed on data sheets as gfs. The gfs of some FET types can be in the A./V. range.

[gingertube]

Rise time in tubes is so damn fast it is never quoted. So what physically is the rise time in a tube?
It is the spread in transit delays for electrons moving from the cathode to the anode - I have never seen an actual value quoted except for photomultiplier tubes. Based upon my experience with them, I would guess that a typical tube risetime would be 1 to 2 nano seconds.
This might be a problem if you are designing an amp to work at 175MHz or above. For audio work its negligable.

Cheers,
Ian