[pminmo]

I have task - rotate platform for satellite antenna with high accuracy and very low speed. There is desired to use near direct drive of platform to avoid mechanic losses and inaccuracy and inertia. NO speed control only position.

Regard,
Andriy

You will need some type of mechanical gearing via gears, screw or timing belt for a reasonable performance with something of that type of mechanical load regardless of motor type.

The 3986 is a better fit with that type of bulky work than cnc.

[yamukha]

Hello Gentelmens.
Thank you for your replays.
Currently I am in China in business trip.
I've got a chance to visit a supermarket of electroniks in Shenzhen and I find something interesting for me. Llook at www.bsjd.com (BAISHAN)
There is stepper motors with toruqe higher then 20 N*m with step angle 1.8 or evev 1.2 grad. IMHO 200 or 300 steps per rotation (2 and 3 phases steppers).
I bought two small motors to check angle and torque performance at small speed coz there are performance characteriscics starting from 100 pulses\s.
Bajshan offers stepper motors drivers povered from 100AC and 220AC as well as 12..40VDC. So I should consult with mechanikers what torque they realy need. And what angle resolution and minimal speed of rotation.
Maybe the simplest solution is to by 2 or 3 phase stepper motor and proper driver form bajshan.
Anyway all conversation about drivers for stepper as well as for
DC\AC converters will be done after my return to home.
Other interesting findings - no need to build 10A drivers.
There are no such powerfull steppers.
Other idea - run in parallel two or more steppers.
Was anybody trying to do it? What results and possible problems?
P.S.
I am very interesting in investigetion and research in field of stepper drivers and DC/AC converters so in the future I am going to chech choppers IC i.e. A3986 (or other similar one) v.s. CPLD/FPGA soluiton with PWM.
What do you think about D-class audio amplifiers for running steppers?
P.P.S
Maybe there are not all answered qwestions.
I goinng to replay them in one or two weeks.
Regards.
Andriy

[amplexus]

There are steppers that run over 10 amps, some of the nema 42's, you can stack two steppers on the same shaft. Proper voltage is 32 x sqrt of inductance. as to a class d amp it might work if you run the stepper as a bldc servo with an encoder. Mariss is currently working on one.

[yamukha]

there is spec for baishn stepper
http://www.bsjd.com/_en/upload/20077556839441.pdf
20 (2832 oz*in) and biggest one 27 N*m (3823 oz*in )
see http://www.unitconversion.org/energy...onversion.html
this greater then nema42. provides near the same 2830 oz*in
http://www.kelinginc.net/KL42H2150-42-8A.pdf
Price for Baishan BS110HB150-06 is 650RNB (near $90) and can be bargained.
IMHO Baishan better by price and performance characteristics.
Maybe do yuo or some body else have other quotes for high torque steppers?
P.S.
Does have Mariss some results with D-class op-amps and steppers?

[Mariss Freimanis]

A rose by any another name is still a rose.:-) If you design audio equipment, you call it a D-class stereo amplifier. You use current mode feedback. If you design step motor drives, you call it a dual full-bridge amplifier. You use current mode feedback. There is no other difference.

Mariss

[yamukha]

Mariss
There are powerfull D class opamps from i.e. Texas Instrument.
I had never tryied to usethem but I have idea to do it.
Real problem is to get samples from TI.
Maybe A3986 too.
I don't know. For me is better reuse something then invent a wheel again.
Regards,
Andriy

[Mariss Freimanis]

I try to reinvent the wheel every day. This way the wheel is exactly as I want it instead of having someone else's idea of what a wheel should be.:-)

Mariss

[yamukha]

Mariss
I was started decade ago in elecrical filed while now working in outsourse of programming of destops. There are different type of thinkig.
I see you thinking like inventor, lot of software developers think like just coders.
I keep in mind engineering.
There was nice cite in CNCzone concerning engineers: trade of optimal solution and efforts\costs (No time to find it).
P.S Sorry I forgot about scientists.
P.P.S sorry for offtop

[amplexus]

Does anyone have a general quadrature decoder schematic that outputs analog signals for velocity, direction shaft angle etc. I know I have seen this somewhere on the web but my searches have only turned up pic solutions. If anyone has a link please post it or the schematic, if not I'll have to take Mariss's advice and attempt to reinvent the wheel.(mine are not always as round as I would like)
Amplexus (Ender)

[amplexus]

yamukha
It seems to me that servo's might be a better solution than trying to use huge steppers. Performance of big steppers is not all that great. Also do you really need that much power. people run mill conversions with nema 23's and have plenty of power to bend and break things.

[ssureshtronics]

Hi all Ers. I had just received the A3986 samples last week, and now i had finished the design in GP board (for A3986 38 Tssop i used a custom PCB). Its working well, and i tried with different motors upto 4A, 48V. When I driving in Microstep I cant get my nedded torque in above 1000 RPM.

Please any one is there to suggest me to get good torque(1.3N-m) even in the 1000 RPM range?. I used Nema 23 1.8deg, 24 VDC, 4A/Phase, Sanyo denki stepper motor.


I'm looking for replacement of my existing design using L297,IR2101,IRF540. Is it good to replace my design with A3986?

Suggestions are welcome

[amplexus]

I need more info about your motor inductance, you should be running bipolar parallel, find max voltage ie 32 x sqrt of inductance and make sure your power supply can supply that, at 5 rpm or more you will get 30% more torque in full step mode. Choose ballscrews of the right pitch, they are more efficient than acme which can rob torque.
Read this entire post, the allegro chips still have issues that they refuse to fix. Watch the pathetic laser pointer video. Search and read Mariss's posts on chopper drives in general. The problems with non synchronous choppers is part of why Mariss did the cpld tutoral. read it it is a much better design than the chipsets you are currently considering. it will get you a gecko drive on a $1 cpld minus a couple of features which you can add your self if you are clever, even as is it is still excellent, If that is still too much hassle buy a gecko, instant fix, more torque dead quiet unkillable.

here is a bit from mariss
> All the chopper type drives are more or less crap Gecko's are quiet
> because they run a synchronous pwm drive. here is a bit of stuff that
> may be helpful
>
> OK, a couple of different subjects so let me take them in turn:
>
> 1) "Microstepping gives less torque." This is a little fiction and a
> little truth.
>
> 1a) Truth: A 100 in-oz motor with a full-step drive gives 100 in-oz
> holding torque. A microstepping drive gives 71 in-oz holding torque on
> the same motor at the same current.
>
> Fiction: You don't buy motors and drives for their holding torque.
> If you did a bolt with a rusted-on nut would be a better bet for the
> +20,000 in-oz of torque it would take to loosen it.
>
> You buy motors and drives because they produce torque while turning.
> Let's see what happens when a full-step drive and a microstep drive
> begins to turn a step motor:
>
> Full-step drive at 5 full-steps/sec on a 100 in-oz motor gives 65 in-
> oz.
> Microstep drive at 5 full-steps/sec on a 100 in-oz motor gives 71 in-
> oz.
>
> What happened to the missing 35 in-oz on the full-step drive? Where
> did it go? Well, it was invested into vibration and resonace to
> shake, rattle and roll as all of you who have the pleasure of using
> full-step or half-step drives know. A full-stepping big motor
> rattles the fillings in your teeth. Score: Microstepping drive 1,
> full-step 0.
>
> 1b) Ok. I'm past the low-speed unpleasantless now. At high speeds my
> full-step drive beats the pants of off XYZ Inc. microstepping drive.
> How come?
>
> How come is because XYZ Inc. is still having their drive
> microstepping at high speeds. That initial 71% torque deficiency
> comes back to haunt them like a ghost from Chirstmas past.
> Microstepping loses all benefit at speeds above 3 revs/sec. It
> becomes a millstone around your neck past 4 revs/sec.
>
> Solution: Simple; stop microstepping above 4 revs/sec. Our drives
> morph from microstepping below 4 revs/sec to full-stepping by 6 revs/
> sec. Morphing is completely transparent to the user; you don't
> notice anything when it happens. You get full-step power at high
> speeds and microstep smoothness at low speeds. You can have your
> cake and it it too.
>
> 1c) I have to mention one other big power-robbing bad motor
> behavior. Mid-band resonance, mid-band instability or parametric
> resonance. You have seen it with XYZ Inc or ABC Inc drives. You get
> your motor up to 5 to 15 revs/sec when you hear a descending low-
> frequency pitch sound from your motors and then they stall for no
> good reason at all a second or two later.
>
> You are the victim of mid-band instability. You try all sorts of
> things. You add friction, you try to remove it. You add load
> inertia, you try to remove it. Nothing helps; the motors growl, then
> they stall for no reason. All you can do is to accelerate past the
> "death zone speeds" and then everything is OK again.
>
> This is a phenomena we understand completely and all our drives
> incorporate the 2nd-order compensation circuitry to completely
> eliminate it. There are no "death zone speeds" with our drives. The
> motors are well behaved at all speeds from zero to 6,000 RPM.
>
> 2) "Full-step is OK, 10 microsteps is better, 1,000,000 microsteps
> is best."
>
> Fiction: The finer the resolution (the higher the microsteps/step)
> the better the performance.
>
> Fact: Step motors run open-loop. They are transducers and like all
> transducers, you depend on the accuracy of the transducer for the
> accuracy of the final result.
>
> Standard step motors have a non-accumulative error of +/-5% of a
> full step. This defines their ultimate accuracy. Think of it this
> way. You have to machine a 1" cube. You have to hold a +/-0.00025"
> tolerance and the cube you machined would be accurate within
> 1:2,000. That's how accurate step motor are.
>
> For a step motor to be that accurate, the drive must be accurate as
> well. It requres a drive accuracy better than 1% (sin/cos
> distortion) and our drives deliver that. A good motor coupled with a
> good drive delivers good results.
>
> Audiophiles will understand this analogy. The best NAD stereo
> amplifier will sound like crap with a pair of $5 computer speakers.
> The best Axium speakers will sound like crap with a computer audio
> amplifier. You put a NAD amplifier together with Axium speakes and
> it's molten sweet smooth honey to your ears.
>
> Same thing here. A 3.6 degree motor from a long forgotten 10MB hard
> drive is the same as a $5 speaker. It is a NEMA-17 piece of crap
> nothing can be done with. It was designed to be a full-step motor
> and no drive will make it be more than that. It is all it was meant
> to be and it is all it will ever be.
>
> Good motors are NEMA-23 to NEMA-34. Some NEMA-17s (Vexta PK245) and
> some NEMA-42s (Bridgport) can be included. Stay away from round 23s
> and 34s, stay away from 'square' 34s with more than 900 in-oz
> holding torque and you will do OK.
>
> Mariss
>
>
> First off, our drives use a synchronous clock oscillator for the PMW
> current servos. This insures no audible beat frequencies are ever
> generated in the operation of the drive, unlike what you get with
> constant off-time chopper drives. You know, the grunting, whistling,
> squealing and hissing that makes that type of drive so endearing.
>
> Our's are totally silent.
>
> The step pulse to switching frequency oscillator phase-locking is very
> simple. It means the PWM oscillator is reset on every step pulse. This
> ensures a 0Hz beat frequecy between the step pulse frequency and the
> PWM frequency at all times. This results again in no audible
> components and more important, no beat frequency components to pump
> the motor into resonance.
> You remember correctly. This behavior occurs when the attack
> > (increase) slope is less than the decay slope and the switching
> cycle
> > period is constant. The result is a subharmonic (multiple switching
> > cycle) waveform.
> >
> > A 'chopper' (constant off-time) circuit is not susceptible because
> > its switching cycle period is not constant; it varies with the
> > current, minimum period at zero current and maximum period at max
> > current. It on the other hand is very susceptible to phase-locking
> > and breaking lock with the other winding's chopper circuit in a step
> > motor drive. This is what causes the familiar hissing, whistling,
> > singing and grunting these drives exhibit.
> >
> > A synchronous (common clock) circuit has a fixed period for both
> > current servo circuits and thus doesn't generate audible phase
> > lock/break subharmonics directly. The overconstrained design does
> > have audible artifacts because of multi-cycle waveforms.
> >
> > Feedforward compensation adds a positive slope component to both the
> > attack and decay slopes enough to insure the attack slope is steeper
> > than the decay slope. This results in a stable waveform that repeats
> > every switching cycle period. No subharmonics are generated and the
> > drive is silent.
>
>
> Maybe 'over-constrained' is not the best term. Having an attack slope
> > less than the decay slope is one constraint. Having a fixed
> switching
> > period is another constraint. Having both in a circuit design
> results
> > in instability because both constraints cannot be satisfied within a
> > single switching period.
> >
> > The circuit adapts to these requirements by delivering the only
> > solution it can, a multi-cycle repetitive waveform. This is
> > undesirable for obvious reasons.
> >
> > To satisfy the third requirement, a repetitive waveform that spans a
> > single switching cycle, one constraint or the other must be removed.
> >
> > If the fixed switching period constraint is removed then you meet
> the
> > third requirement, a stable waveform. The most common solution is
> > a 'constant off-time' current servo. This works OK so long as there
> > is only a single servo circuit. It is the problems introduced by
> > needing two such servos for a step motor drive that makes removing
> > this constraint undesirable.
> >
> > The other choice is removing the attack/decay slope constraint. An
> > effective technique is to add a constant positive slope to the
> > attack/decay waveform. The sum increases the attack slope and
> > decreases the decay slope. A stable waveform results when the attack
> > slope becomes steeper than the decay slope.
> A synchronous (common clock) circuit has a fixed period for both
> >> > current servo circuits and thus doesn't generate audible phase
> >> > lock/break subharmonics directly. The overconstrained design does
> >> > have audible artifacts because of multi-cycle waveforms.
> >> >
> >> > Feedforward compensation adds a positive slope component to both
> > the
> >> > attack and decay slopes enough to insure the attack slope is
> > steeper
> >> > than the decay slope. This results in a stable waveform that
> > repeats
> >> > every switching cycle period. No subharmonics are generated and
> > the
> >> > drive is silent.
> It's simplicity itself. Form a summing node using two resistors. One
> resistor is driven by the slow-attack/fast-decay waveform. The other
> resistor is driven with a ramp waveform synchronous with the switching
> cycle.
>
> The summing node will display the sum of these waveforms; the slopes
> add (+,+) during attack, subtract (+,-) during decay. Scale the
> resistors to insure the summed attack slope is steeper than the
> differenced decay slope.
> The midband resonance circuit adds a derivative component (+80 degrees
> phase lead) to the system phase angle to eliminate parametric
> resonance. It involves sensing the rate of motor load change
> (derivative of motor torque) which then phase modulates the internal
> step pulse timing.
>
> The sine/cosine to quadrature reference morphing extracts the maximum
> possible torque from the motor at speeds above where microstepping is
> of any further benefit. It is an "area under the curve" thing. The
> sine function has only 78% the "area" of a squarewave (full-stepping).
> The drive adapts by morphing into a full-step drive at higher speeds.
> All that is in addition to the second-order damping circuitry used for
> > >midband resonance (parametric resonance) suppression. Another
> > >enhancement is morphing the sine/cosine microstepping reference
> into a
> > >quadrature squarewave (full-step) reference beginning at 4 revs/sec
> > >and finishing at 6 revs/sec
> phase sensing can be done in (at least) two diferent ways. You can
> model the
> behaviour of the motor as a series connection of an inductance and a
> resistor. In the frequency range where mid band resonance occurs the
> inductance is dominant. The drive acts as a voltage source resulting
> in the
> microstepping sine waveform to degenerate to a square wave and the
> current
> to a triangular waveform.
>
> Phase can be measured by detection of the zero crossing of the
> current with
> a capture timer and correlating that to the voltage waveform. Another
> approach would be measuring the average absolute current. As the
> current
> through the two coils are triangular shaped and 90° phase shifted,
> adding
> their absolute values gives nearly a flat curve. Since voltage and
> inductance are constant you can calculate the ohmic resistance from
> it which
> represents copper loss and mechanical load. As mechanical load
> varies due to
> mid band oscillations the average current also does. The first
> method is
> best for a microprocessor the second would require lots of math but
> can be
> easily implemented in a little analogue circuit. That was the
> sensing part.
>
> The second part is phase correction. There's only one way to do this
> because
> the drive is in constant voltage mode and you do not have any
> influence on
> the amplitude. So all you can do is to delay the step pulses to
> shift the
> phase. As Jeff said, this can be easily done with an analogue
> sawtooth ramp
> generator, at least as long as you microstep resolution is not too
> high (say
> 4..16). For higher factors you'd have to completely swallow pulses,
> remember
> them and insert the exact number later. This would require complex
> sequential logic or a fast microprocessor.
>
> The third part is to connect the sensing and correction to close the
> control
> loop. You have to insert a filter with the correct frequency and phase
> response between the two.
>
>
> Midband resonance results from a system phase lag of 180 degrees.
> > > The step motor has an inherent 90 degree lag between torque and
> > > position.
> > >
> > > Everything works OK while the drive acts as a current source (0
> > > degrees phase lag). As speed increases, the drive must eventually
> > > revert to a voltage source. This adds another 90 degrees phase
> lag
> > > for a total of 180 degrees and this is a setup for unconstrained
> > > oscillation.
> > >
> > > Our drives add a phase-lead component of about 80 degrees to the
> > > system loop phase lag. This is a derivative component,
> specifically
> > > a rate of load change sense. This makes the drive immune to
> midband
> > > instability.
>
>
> Low-speed resonances are nulled by a trimpot. It compensates for a V/L
> offset error all drives are prone to. It results in a complete
> suppression of all low-speed (<1 rev/sec) vibration.
>
> A little bit of reflection would give you the answer. Can a half-
> step drive hold
> position at the half-step location? The answer is yes.
>
> A half-step drive is really a 2 microstep drive and is fundamentally
> the same as
> any resolution microstep drive.
>
> A good step motor's torque is the vector sum of its phase currents.
> If you want
> the torque to be constant (you do) and independent of angular
> position, use sine
> and cosine winding currents (and we do). This is because the
> trigonometric
> identity '1 = sin^2 + cosine^2' gives a constant (1) for any angle.
> Torque is
> then independent of shaft angle and it doesn't matter where the
> motor is
> stopped.
>
> On a related topic. A step motor's shaft angle also depends on the
> torque
> applied to the motor. The relationship is sinusoidal; applying a 10%
> of holding
> torque load displaces the shaft 0.115 degrees, 50% moves it 0.600
> degrees and
> 90% gives 1.283 degrees. Release the load and the shaft moves back
> to the
> original rest position.
>
> Angle = full-step angle (arcsin (load torque / holding torque) / 90)
>
> This relationship is true for full-step drives as well as
> microstepping drives.
> Apply 100% of holding torque load and the shaft will move from the
> rest position
> by 1 full step. Apply 101% and the motor cannot hold any position.
>
> Where this matters for microstep drives is if you are expecting to
> hold a
> position angle equal to the microstep resolution. In other words,
> how much
> torque can you apply to the motor before you are out of position by
> 1 microstep?
> We build 10-microstep drives so the answer is "T = holding torque *
> sin 9
> degrees" or 15.6% of holding torque. Again, when the load is
> released, the motor
> returns to its original location.
> ) Microstepping requires accurate sine and cosine currents. The
> sin/cos look-up table is in the CPLD. The current set resistor is the
> bottom half of a voltage divider. The voltage formed is the analog
> multiplier for the sin/cos DACs that create the sin/cos references for
> the drive.

[herdmay1]

After stewing a number of feature issues, just sent my prototype out. :banana:

i am posting for the first time and hence a hello to all.
also i learnt so much from here that i really have to thank everyone posting.

i have just build a working a3986 board and would like every one to see it and comment, and also wish and hope that it is useful to someone seeking to know more and build one.

i have the pcb layout/schematic etc but i dont know how to post images..plz help

RK

[jalessi]

Herdmay1,

1: You will need to zip up the files in order to post them.
2: Click the manage attachment button.
3: Click the browse button , locate your file and upload it.

See attached link for the zip program, it is free.

http://www.7-zip.org/

See attached pictures for help.

Jeff...

[herdmay1]

Herdmay1,

1: You will need to zip up the files in order to post them.
2: Click the manage attachment button.
3: Click the browse button , locate your file and upload it.

See attached link for the zip program, it is free.

http://www.7-zip.org/

See attached pictures for help.

Jeff...

thnx! jeff

here i am uploading the images of the layout

[MechanoMan]

As a refresher to those who haven't waded through 60+ pages of discussion:
The A3986 has a huge problem in that there's a large fixed "dead time" which creates a large minimum duty cycle.

This makes it impossible to resolve the lower currents of microsteps. This causes minor static position error, and major dynamic distortion when moving which will make it very vulnerable to stepper stalls.

With higher voltage supplies, the threshold of the lowest current it can create increases, making more levels of microstepping impossible to create. This apparently wasn't intended to use a high voltage supply, but the low voltage ones are far too slow to run CNC equipment.

It might be a good idea to just try to turn off microstepping, but honestly, nothing's gonna make this chip work properly for CNC equipment IMHO.

[herdmay1]

As a refresher to those who haven't waded through 60+ pages of discussion:
The A3986 has a huge problem in that there's a large fixed "dead time" which creates a large minimum duty cycle.

This makes it impossible to resolve the lower currents of microsteps. This causes minor static position error, and major dynamic distortion when moving which will make it very vulnerable to stepper stalls.

With higher voltage supplies, the threshold of the lowest current it can create increases, making more levels of microstepping impossible to create. This apparently wasn't intended to use a high voltage supply, but the low voltage ones are far too slow to run CNC equipment.

It might be a good idea to just try to turn off microstepping, but honestly, nothing's gonna make this chip work properly for CNC equipment IMHO.

very true!

the point however is that A3986 can very well be used by people who need microstepping with reasonably less effort. and trust me..there are applications which suit A3986 and vice-a-versa

kindly do not fall for claims in the datasheet (30-500Watts)and like mechanoman pointed out, certainly not for current gobbling , critical apps like CNC

make sure you assume the below before starting and u willbe happy using a3986

1. do not use voltages above 30v
2. do not use mosfets other than irfz24 ( or with lower VDS and lower Ciss)
3. u cannot draw more than 2 amps (even if you short the rsense)
4. current will keep on decreasing as the speed increases and nothing can be done to avoid this so...follow the torque chart for standard motors.


i hope the above helps

[ratep2001]

Hi,

First time on this forum, so I hope I don't ask something which has already been discussed, although I already searched the forum as good as I can.

I could see that an issue with Enable pin on A3986 was already discussed, and I downloaded a white paper about PWMing the Enable pin. However, my problem has a different nature.

I want to disable the step motor i.e. the A3986 driver between sequences of steps, for example if the user leaves the machine overnight or during longer period of inactivity. Each new sequence is started from zero position, determined by limit switches and thus I don't rely on powered step motors to keep a defined position from a previous sequence of steps, so I can just disable the A3986.

But here is the catch - if I disable the A3986, I canot start it again, there are no pulses coming out of the driver (I don't know if it is just me, a bad IC or it is a general problem).

I think that the reason is in the Undervoltage lockout of the bootstrap capacitor power supply for the high side gate drivers after a period of inactivity (no switching on of lower FETs). Since I don't know how long the period of inactivity is, I cannot simply put a bigger bootstrap cap - I will need a dedicated power supply, probably just a simple bleeder resistor and a zener diode across cap.

Anybody on the forum experiencing the same problem and having an elegant solution?

Regards,
P.

[eSilviu]

Hi,


But here is the catch - if I disable the A3986, I canot start it again, there are no pulses coming out of the driver (I don't know if it is just me, a bad IC or it is a general problem).

Regards,
P.

Why don't you use curent setup to make an "fake" disable?:
I would use Vref to set:
- 2A for normal operation
- 0.5A if driver is unused for 1 minute or so
This way motors won't get hot after a long unused period, with driver powered. And the microstep position is remembered too, so this "enable" can be used on the fly, when machinning, to lower motor curent when not used.

[herdmay1]

Hi,

First time on this forum, so I hope I don't ask something which has already been discussed, although I already searched the forum as good as I can.

I could see that an issue with Enable pin on A3986 was already discussed, and I downloaded a white paper about PWMing the Enable pin. However, my problem has a different nature.

I want to disable the step motor i.e. the A3986 driver between sequences of steps, for example if the user leaves the machine overnight or during longer period of inactivity. Each new sequence is started from zero position, determined by limit switches and thus I don't rely on powered step motors to keep a defined position from a previous sequence of steps, so I can just disable the A3986.

But here is the catch - if I disable the A3986, I canot start it again, there are no pulses coming out of the driver (I don't know if it is just me, a bad IC or it is a general problem).

I think that the reason is in the Undervoltage lockout of the bootstrap capacitor power supply for the high side gate drivers after a period of inactivity (no switching on of lower FETs). Since I don't know how long the period of inactivity is, I cannot simply put a bigger bootstrap cap - I will need a dedicated power supply, probably just a simple bleeder resistor and a zener diode across cap.

Anybody on the forum experiencing the same problem and having an elegant solution?

Regards,
P.

try gving a short pulse to reset along with the pull down of enable!